Building Earth-like Planets: from gas and dust to ocean worlds. - Episode Artwork
Culture

Building Earth-like Planets: from gas and dust to ocean worlds.

In this episode, we explore the fascinating process of building Earth-like planets from dust and gas, focusing on the essential role of water in creating habitable ocean worlds. Join us as we delve in...

Building Earth-like Planets: from gas and dust to ocean worlds.
Building Earth-like Planets: from gas and dust to ocean worlds.
Culture • 0:00 / 0:00

Interactive Transcript

spk_0 So first, a very heartfelt thank you to Mr. and Mrs.
spk_0 LaBonna Frostowski for this honor.
spk_0 I'm really pleased to be here to give the first of these lectures.
spk_0 And also a tremendous thanks to Oxford as a whole for the warm welcome I've had this week.
spk_0 And particularly to Tamsen-Mayder for leading the charge to get me here as the astrophelow
spk_0 for this year at Oxford.
spk_0 I'm really very honored and it's been such a pleasure.
spk_0 And I've learned so much since I arrived and certainly my knowledge of noble gases has
spk_0 gone up by a thousand fold, which is a great thing actually very very useful.
spk_0 So thank you all very much indeed.
spk_0 So this is what I'll talk about today indeed.
spk_0 Building Earth-like planets from dust and gas to Earth-like ocean worlds.
spk_0 And the photograph I'm showing here is really in honor of Mr. LaBonna Frostowski.
spk_0 This is the Angara River in Russia.
spk_0 And the reason I'm showing it is because it's a beautiful image of water and life and rocks.
spk_0 And that's really what I'm going to try to talk about.
spk_0 So I'm always telling my postdocs and I've also generously told your postdocs here this week
spk_0 that they should be trying to answer the largest question that they can.
spk_0 And of course as scientists we can only meaningful and remake incremental steps.
spk_0 But we need to be aiming towards some large question.
spk_0 And so the large question that I like to say that I'm working on is are we alone?
spk_0 That's a very fundamental question of humankind.
spk_0 And so to begin to make progress toward that as a planetary scientist
spk_0 I ask myself what are the things that we know with a certainty are required by every kind of life we've discovered so far?
spk_0 And one of them is water.
spk_0 And so I'm going to take a line from NASA's line of book of quotations and say we're going to follow the water today.
spk_0 And what I'm hoping to convince you of by the end of this talk is that it is possible indeed likely
spk_0 for any rocky planet to accrete with enough water that it should have oceans almost immediately
spk_0 and become habitable in very short order without any stochastic by chance addition of water later on.
spk_0 And so what I'm going to try to do is track the water through the accretion of planets from the smallest particles
spk_0 up to planetary size objects talking about what I consider to be the critical bits of chemistry and physics along the way
spk_0 that allow them to retain water such that they could have oceans almost immediately.
spk_0 That's the goal.
spk_0 So the first question might be where are we in time?
spk_0 I always like to think about things in terms of time.
spk_0 So here's a cartoon of our solar system's development from on the left hand side 4.568 billion years ago.
spk_0 And I hope that that's a number that for many of you looms large in your mind.
spk_0 And if not then for the rest of you I hope that you'll remember it forever after this lecture.
spk_0 I used to tell my undergraduates at MIT that if they could not recite to me 4.568 they failed.
spk_0 And so that there will be no exam I promise.
spk_0 But that is the age of the very first solids that condensed out of the dust and gas cloud that became our planets.
spk_0 And so that's our beginning point.
spk_0 And sometimes people say that's the age of the earth but of course that's absolutely not true.
spk_0 And I hope you'll understand the ways that that's not true if you don't grasp that already.
spk_0 That's really just the first little pebble size object that condensed in our solar system.
spk_0 And so you can see very very shortly after that something that I'm terming here the moon forming impact.
spk_0 So this is the last giant accretionary impact that built our earth.
spk_0 And we think also through off enough debris to create the moon.
spk_0 And certainly there were many impacts to the earth after that but none of that magnitude.
spk_0 So that's kind of where our earth begins in a sense with the moon forming impact.
spk_0 Very shortly afterwards the first evidence of water oceans on the earth very very early.
spk_0 And then there's a sort of a late heavy bombardment that you see in cartoon.
spk_0 And then there's the gradual development of tiny life.
spk_0 And then finally the rise in multi-cellular life and the rise in oxygen.
spk_0 And then extinctions that of animals large enough that we can easily recognize this in the fossil record.
spk_0 And so at the other point I want to make with this slide before going on to zero in on the bit of time we're going to talk about today.
spk_0 It's that as we know and of course at Oxford and indeed and other universities there's an unbelievable heritage of the work that's been done on ages in that last bit during those extinctions when we have lots of fossils.
spk_0 And we've really been able to finally tune our understanding of the geochronology in the last say 500 million years.
spk_0 We understand a lot about the rate at which things happen and exactly when they happened.
spk_0 And then there's sort of what several of my friends and I call the boring billions in the middle where really we actually don't have as many landmarks as we might.
spk_0 And we don't know as much about rates and processes and events.
spk_0 And the other thing that I hope to impress upon you if you didn't already think this way is that just as we have great fidelity in our timeline very recently we actually have tremendous fidelity in our timeline way back where we're going to be talking about right now.
spk_0 And that's I think quite a remarkable achievement for science.
spk_0 So I'm going to show you next slide that just goes from 4.568 just to the moon forming impact and a little bit beyond through to the ocean.
spk_0 And now it's now it's vertical. I hope you'll forgive me for this.
spk_0 This is the very bottom of the geological timeline from Martin Van Crane a dog.
spk_0 And you can see on the right at the bottom 4.568 4.568 billion years the first solids in the protoplanetary disk.
spk_0 And then right after that 4.404 earth's oldest crustal material.
spk_0 And these are minerals called zircons that have weathered out of sedimentary rocks and have been found in western Australia.
spk_0 And they date to as oldest 4.404. So first let's look at 4.568.
spk_0 This is a calcium aluminum inclusion writ large. It's actually about a half of a millimeter across.
spk_0 And it is an example of the earliest solids to form in our protoplanetary disk out of dust and gas.
spk_0 And they can be dated quite accurately.
spk_0 And they come primarily the oldest ones come from a meteorite called Ayende that fell in Mexico in the 1960s.
spk_0 And happens to hold some of the very oldest of these calcium aluminum inclusions.
spk_0 They're virtually tiny pebbles. Some of them are centimetric in size.
spk_0 And they have the oldest ages of anything that we found.
spk_0 And so we have them because fragments of this unprocessed material falls to earth as meteorites.
spk_0 And it's really just the record of these earliest times.
spk_0 And so this is obviously not in real color. I wish they really were colored like this.
spk_0 But they're colored according to the composition. It's magnesium and titanium and calcium that have been colored in this slide.
spk_0 And the things that magnesium, titanium and calcium have in common is that they condense at the highest temperatures.
spk_0 So as the gas and dust is cooling, these are the first things to condense out. And they give us that earliest age.
spk_0 Now here are the zircons that give the next age 4.404 billion years ago.
spk_0 And they are from the Jack Hills in Australia. And you can see the scale bar. They are 100 microns.
spk_0 So that's a tenth of a millimeter scale bar. And I think this is, I'm not sure is this cathode luminescence or this is some sort of back scattered electron image.
spk_0 But what it's showing you is composition in color. And so you can see how the compositions of these zircons have changed over time as they formed.
spk_0 And then you can see the many, many holes where they were individual ages were obtained. And some of these ages are as old as 4.4.
spk_0 And so for those of you who have not actually experienced the wonder that is a zircon, it's zirconium and silicon and oxygen.
spk_0 And they like very much to crystallize with uranium in their matrix. And uranium decays to lead and gives us a radiometric way to measure their age.
spk_0 And the reason they're so good at this is because they will accept the uranium into their crystal. But they will reject any lead in their community.
spk_0 And so all the lead that ends up in them is from the decay of their uranium. So it's a very high fidelity clock. And they're also very beautiful.
spk_0 And so I will tell you that the earrings I'm wearing are natural zircon earrings from India. And I hope they're not very radiogenic.
spk_0 But they're quite a wonderful mineral. And so these have given us the oldest ages of an earth created rock that we have. And there is none older for many reasons the crust in between has been destroyed.
spk_0 And so that gives us this span of time. 164 million years to go from pebbles to an earth that has oceans.
spk_0 Because some of these zircons record the presence of liquid water on the surface of the earth as they were forming in their oxygen isotopic composition.
spk_0 And so that is a very short amount of time. And that's the amount of time that we're going to traverse in the next 35 minutes or so.
spk_0 All right. So here's a cartoon of the disc of dust and gas that I mentioned in my title from which are condensing these little fragments.
spk_0 And out of which are then growing pebbles and boulders and then what we call planetesimals. These miniature planets and tens to hundreds, maybe a thousand kilometers and radius growing larger into planetary embryos and finally all the way into planets.
spk_0 And that's the series we're going to traverse. But what is the material that we're starting with? I'd like to talk about this because it points out two things.
spk_0 One is one of the fundamental differences between physics and earth science and that is the use of the word metal, which is used entirely different in the two fields, differently in the two fields.
spk_0 All right. So this is a table from Anna Lauder's paper in 2003 in which she is looking at the composition of the sun.
spk_0 The composition of the sun is, I think, a probably very accurate proxy for the composition of the bulk material that made all of our solar system.
spk_0 And so it's very interesting to see what the sun is made of. And I did not know this until I came back onto the faculty at MIT and started working with astrophysicists that the sun and other stars are measured in terms of three components, X, Y and Z, which are hydrogen, helium, and everything else.
spk_0 And so this is how much of everything else that there is. And they're all called metals. If it's not hydrogen and helium, it's a metal, which is different from our use in earth science.
spk_0 But what is the point of this? Why am I showing this? Because we only have between 1 and 2% of everything else to make the planets out of.
spk_0 And we do have plenty of hydrogen, but not as much helium on our earth. But the vast majority of what the earth is made of and what you and I are made of and what is required to create light is in that Z. It's in that other category.
spk_0 So when we try to trace what happens to water, which is only the tiniest part of the Z, which is already the tiniest part of the whole, it becomes a really different exercise, difficult exercise, because you're just tracing the trace of the trace through these processes.
spk_0 Because most everything is hydrogen and helium. And the rest of the Z is almost everything else. So it's just a trace of a trace that allows a planet to be habitable.
spk_0 And so the point of this is that it only takes a trace of a trace to make oceans of water. And that's really the message in the end.
spk_0 So one pointed out to me once that if you look over the year of publication from 1984 to 2003, the amount of Z in our Sun is dropping precipitously.
spk_0 But of course that's a measurement error, not measurement error, but a changing in the accuracy in the techniques of measurement, not in an actual change.
spk_0 So what happens to this material? So on the left in this cartoon that I've drawn, we have the little pebbles and bits of dust and the things that are creating in the disc.
spk_0 And what is happening is they form into these bodies in the arc in the middle called planetesimals.
spk_0 And you can think of a planetesimal as an asteroid. And I think that some asteroids probably really were planetesimals. And many of them are just fragments of planetesimals.
spk_0 Vesta, the asteroid Vesta is a beautiful example of a planetesimal. So these are bodies tens to hundreds of kilometers in radius that are then going to gravitationally interact and build up into planets.
spk_0 And I want to make an important distinction here, which for some of you is second nature and you understand exactly what I'm drawing.
spk_0 And for others of you is maybe a new idea. And that is the topic of differentiation.
spk_0 Now it's a critical, critical step in planets. And that's when you take the metal part of what's building up and the silicate part of what's building up.
spk_0 And you allow the metal to sink into the middle and make a metallic core surrounded by silicate mantle, which is the structure of our earth and the structure of mercury and Venus and the moon and Mars.
spk_0 But not the structure of the original, caundritic, undifferentiated primitive, those are all adjectives that apply to the same class of materials, undifferentiated primitive material in which the metal and silicate are mixed on an intimate scale.
spk_0 So you take a lot of that intimately mixed metal and silicate and you put them together.
spk_0 And the first thing that happens on the left of this arc is they begin to heat up. And that's primarily because of the decay of a short-lived radioisotope of aluminum that really is only around in the first one and a half million years of our 164 million year window of opportunity.
spk_0 But it causes these planetesonals to be hot. And if they heat up sufficiently, the metal fraction leaks into the metal and forms the core.
spk_0 And so we think that there were planetesonals like on the left that are heated in some onion skin way, but not actually differentiated into a core and a mantle.
spk_0 And then others, you can see with the red hot metallic core, have differentiated. Some have retained a caundritic primitive crust and others have melted all the way through.
spk_0 Now, why is this important? Because this is the first important step where we need to trace what might happen to the water.
spk_0 So you have some amount of water in this dust that's creating into planetesonals. But if the planetesonals are heating up to 900 or 1,200 degrees Celsius or in fact much hotter if they're allowed to because of this aluminum 26, how can they possibly retain water?
spk_0 And so that's the first question that we have that I'll show you the beginnings of an answer in a moment.
spk_0 What happens to these planetesonals is many of them are swept up and become part of planets and some of them are broken apart and later delivered to us as meteorites.
spk_0 And so that's the far side. That's the library of material we have from which we can imagine to make planets.
spk_0 And so here is a graph of data on this library of meteorites that have fallen to the earth that are really just a selection of some subset of the material that was making planets early in the solar system.
spk_0 So here's a graph that shows you water in weight percent from 0 to 18 weight percent on the vertical axis and carbon from 0 to 4.5 on the horizontal axis.
spk_0 So here are two critical things water and carbon. And these are bulk compositions of these meteorite samples of many different kinds, both the primitive undifferentiated and the differentiated processed types of meteorites are shown here.
spk_0 And of course there's some you see that go up to almost 18 or almost 20 weight percent of water. And let me tell you that as many orders of magnitude more than you need to make a planet with oceans.
spk_0 So that's more water. In fact, Kevin's only always says the problem is getting rid of water, not adding water. And I certainly have come around to that opinion as well.
spk_0 So looking back here, the very wettest ones are the ones on the left that are these onion skin undifferentiated ones.
spk_0 And the ones that have differentiated are drier, but they're not perfectly dry. And so how would a body like a planet test will hold on to water so that it would own some water as it went to a creed into planets.
spk_0 It would still bring water to those planets. So I've been doing quite a bit of work on this in the last few years. And this is a paper that is being led by Roger Foo, who's, he's just a third year graduate student at MIT.
spk_0 He's mostly working with Ben Weiss and we did this project on the side. And he's someone to watch for. He is very, very bright guy. He's mostly doing paleo magnetism. He's a person who just recently measured HED meteorites from Vesta and showed that Vesta had a core dynamo in a magnetic field.
spk_0 So he's doing some quite interesting work. And it is copious spare time as a PhD candidate. He did this project with me on the side. And our question was, you start on the left here with a large early accruing planetesimal.
spk_0 And with progressive radiogenic heating from the left to the right, the first thing that's going to happen is any isses or water bearing minerals are going to be heated to the point that they break down and they give off their water as a free fluid.
spk_0 And this free fluid is going to start to percolate upward through buoyant flow in the planetesimal. And Ed Young at UCLA has done a lot of work on this. And he said that this percolation in bodies less than about 40 kilometers in radius is not very efficient.
spk_0 And actually that will get recirculated and the mineral grains will rehydrate and so that body will hold onto its water. But in bodies larger than 60 kilometers, the fluid loss is relatively rapid.
spk_0 But then these fluids may freeze in the extremely cold boundary layer against space near the surface. So you would end up with a body that was dry on the inside and wet on the outside.
spk_0 And so then we've asked the further question, if the silicate part starts to melt, what's going to happen with that melt? Is the melt going to erupt onto the surface? Is it going to remove all the remaining volatile to space? Or can in fact this planet has to retain both a wet crust and it silicates in the interior?
spk_0 And so what Roger and I have done is we have calculated the density of the melts that would be expected from many different classes of meteorites. And we found that some of those melts are buoyant and they should erupt on the surface, allow volatiles to go free into space and dry their parent body.
spk_0 But others are actually denser than their coexisting solid and they would be expected to stay inside and you would keep a planet decimal with a bit of a wet crust.
spk_0 And so these are the kind of processes we're trying to follow through a planetesimal evolution. I actually think this is a really interesting field and I have a lot of ideas for more work on planetesimals.
spk_0 There's so critical to understanding what kind of planet you can build. You have to make a planetesimal first. And in fact at my department on Monday and Tuesday we're having a big workshop on planetesimals. We have about 65 people coming from Europe and Asia and Australia and all over the US to come talk about this problem.
spk_0 So the question is how do you form a core on a little body like this? How does melting proceed? What kind of volcanism could you have and can you keep hold of the water? And the reason it's so interesting is because first of all there's very low gravity and the melting process is entirely unlike anything that happens on earth.
spk_0 So for those of you here who like to study earth melting processes I'll just talk about this for a moment. I think it's really interesting.
spk_0 So on the earth melting happens in several ways. One is by having pressure released from it. If you have packages of the mantle that are rising up toward the surface, the pressure on it goes down and down and it may then be able to melt through the release of pressure.
spk_0 And another way that melting happens on earth is by the addition of water in a subduction zone. You add water to the hot mantle and it melts because the water lowers its melting temperature.
spk_0 Neither of those things happen on planetesimals. What happens on planetesimals is they are heated through their internal heat, through radiogenic decay.
spk_0 And so the heat is hottest in the middle of the body and the least on the outside. The melting starts at the bottom of the mantle instead of at the top of the mantle.
spk_0 And because heating begins progressively from the temperature of the disc up to the melting temperature of the silicates, they are able to pass through the different thermal regimes on their way, lose their water, for example, dewater all their hydros minerals and become a dry melting source before they melt as silicates.
spk_0 So for those of you who like to study melting I think that's quite an interesting regime and not one we're used to.
spk_0 But the point of this slide is just that we think that there are different ways that even fully differentiated planetesimals could hold onto some water but it would be heterogeneous in that body.
spk_0 And so here's the same part of the library of meteorite samples to see if we might be right or wrong. Here's water on the vertical axis as before but only to 1 weight percent, not to 18 weight percent.
spk_0 And on the horizontal axis carbon only up to 0.5 weight percent. So we're looking at the extreme corner of the previous slide that I showed you right down near the origin.
spk_0 And the points that I've highlighted in orange are samples from differentiated planetesimals, ones that have been through this core forming process, ones that have been through some degree of melting and might be expected to be fully dry but in fact they're not fully dry.
spk_0 There's one out there up there at about 5 or 6,000 parts per million of water that is 0.5 or 0.6 weight percent. But even the ones down near the origin are in the vicinity of 0.1 weight percent of water which is 1,000 parts per million in its bulk composition.
spk_0 So 1,000 parts per million that doesn't sound like very much but I'm sure many of you realize that if you were to take all the water on the surface of the earth and mix it back into our silicate mantle.
spk_0 It's only between 200 and 250 parts per million because that's the equivalent of four oceans right there on a sample that's fairly dry.
spk_0 So that's a part of my reasoning for saying that planetesimals retain water as they're going into the planet building process. And so we have this picture like this.
spk_0 So we have planetesimals and they some are differentiated, some are not differentiated. They are colliding together and forming embryos which are these larger bodies and you can see I'm showing on the embryo that some of these big impacts actually melt a portion of the embryo.
spk_0 And now we have a magma ocean on the surface of the embryo. This is a miniature planet that's say 1,000 or 2,000 kilometers in radius.
spk_0 Some people say Mars is an embryo that was stranded without anyone to continue colliding into it and making it into a proper size planet. So it's sort of an embryo it's a little embarrassing for Mars I think.
spk_0 And then these embryos collide together and make planets. And so here's a summary of where we are so far.
spk_0 So on the left we have the calcium aluminum inclusion. Oh sorry, on the very far less we have gas. This is just the gas disc around our growing young star.
spk_0 And then there's the calcium aluminum inclusion at 4.568 billion years ago. And on the right you see the planetesimals and the embryos growing up into planets in the middle you see an asterisk which is the universal scientific symbol for then something magical happens.
spk_0 Because I don't know if any of you are working on this. I didn't happen to talk to anyone who was but there are a few people trying to bridge that gap.
spk_0 And here's the problem. If you have got dust and tiny, tiny pebbles and they have a little bit of a differential velocity but they're colliding together in that spinning disc.
spk_0 They can stick partly through electrostatic forces. I mean it's easy to say it's very much like the dust bunny under your bed at home. These things kind of clump together naturally.
spk_0 And when you get up to things that are hundreds of kilometers or tens of kilometers in radius gravity is now important and they can actually collide in a creek.
spk_0 But for those middle-sized bodies, electrostatic forces are not strong enough and neither is gravity. And so there's really sort of an open question. How do you get from the dust to the planetesimals?
spk_0 So we know for a fact that it happens because here we are. And so there are a lot of ideas having to do with for example, kelvin Helmholtz instabilities and turbulent compaction. But that's a really wonderful unsolved problem.
spk_0 So now to skip on to the accretion part. Here's a beautiful figure from Raymond in 2006. And it's illustrative of many models that have been made of the accretion part of planet building, the gravitationally driven accretion.
spk_0 So let me explain what we're looking at and why. What they've done is they've made a dynamical code that imagines something like a thousand and 54 planetesimals going around the sun. And the code tracks every single one.
spk_0 And when they have a certain proximity and relative velocity, they're soon to have accreted into a planet. But sometimes they're not quite that close. And so they might gravitationally bend each other's orbits and cause planetesimals to mix in or out or cause planetesimals.
spk_0 Orbits to become less round that is more eccentric. And so the question is the question we're trying to solve at this point is how much mixing is there radially when you're making a planet? Traditionally it was thought that the earth accreted from material at the radius of the earth. And Venus accreted from material at the radius of Venus.
spk_0 And now we know neither of those things is true. Not only do planets accreted from material at a wide range of radii, but the planets themselves probably migrate since they were made. And so it becomes more complicated. Why is this important before we walk through this? Because ostensibly, and there seems to be good evidence for this, the farther away you get from the sun, the wetter your little planetesimals are, the more ice.
spk_0 And so if you want mercury to start with some water, it would be good to have some mixing inward while it's accreting. And so the question is how much mixing is there? So there are six time slices here. And we'll start at the top left at zero million years. And the vertical axis is eccentricity that is how far out of round is the orbit. And the horizontal axis is semi major axis. And so you're going out to just, I guess, five AU. And they do have a Jupiter put in out there past the edge of the planet.
spk_0 So if you're looking at the surface of the screen, so to speak, because those gravitational perturbations are very important. And then the color scale is water. How much water? So you can see in the first slide you start with dry things near the sun and wetter things and then wetter things going further out. And you can see to the right at 0.1 million years. There's been a lot of gravitational interaction and a lot of the orbits have become quite eccentric. And some mixing radially is occurring. And as you step through, you see more and more radial mixing of material as orbits are perturbed by neighboring energy. And so as you see, I'm going to have a little bit of a sense of how the
spk_0 planet has moles of material from far out in the disk is thrown inward and
spk_0 material from inward is thrown outward.
spk_0 Until finally in the last disk you see that this particular run has made four
spk_0 planets, each one of which is wet. There's no planet that's all red.
spk_0 Even the innermost planet has had material mixed into it.
spk_0 And so I think that you can just never claim to make a perfectly dry mercury by
spk_0 only accruing material from next to mercury.
spk_0 There's so much efficient radial mixing.
spk_0 And I could have showed you a paper from 2013 or 2005 and from
spk_0 a number of different research groups they seem pretty in agreement about this.
spk_0 So if we think that we can retain some water and even relatively dry planet
spk_0 decimals and then even mixing of radial mixing of planet decimals adds water
spk_0 to planets. So then we're I think upping the chances that these planets end up
spk_0 with water. And so what happens? Now here's where we get to the stuff that I love
spk_0 the best. What happens when you have an embryo, a big planetesimal,
spk_0 coming in and smacking into another embryo to make a proper size planet.
spk_0 And so this is a cartoon of this process made by Casey Liss at the University of
spk_0 Maryland. And so here's this innocent hapless embryo going around a star.
spk_0 Unbeknownst to it being pursued mercilessly by this other embryo.
spk_0 Of course this is just an artist's image of what happens, but I like it because
spk_0 it's visceral. And so what happens is the energy of accretion is often sufficient
spk_0 to melt the entire target body to some depth. And you end up with a magma ocean,
spk_0 which is one of my very favorite things to study. And also I think I think really
spk_0 very poetic, isn't it? To think of a planet during its growth, having so much energy
spk_0 input that it should become just a ball of liquid magma in space. And so there are
spk_0 many dynamical simulations. I could have showed you computer models of this instead
spk_0 of this cartoon that with careful energy balance do indicate that for example in the
spk_0 moon forming impact for the earth that that impact would have been sufficiently energetic
spk_0 to melt the earth to between 2,000 kilometers depth or all the way down to the core.
spk_0 And so then the earth would have started again as a magma ocean. And so starting, I don't
spk_0 know 10 or 15 years ago or something, I started to ask myself if I want to make progress
spk_0 in understanding how terrestrial planets form. I need some kind of good starting point
spk_0 for a forward model. Everything I've showed you up to this point, you can model bits of
spk_0 it and we begin to get a hold of it. But it's hard, it's very stochastic or it's sort
spk_0 of unknown conditions. We don't really know how things solidify on planetesimals. But
spk_0 this is a clean starting condition. We know a lot about how silicate melt solidifies.
spk_0 Maybe not enough, we don't know enough, but we know a lot. And so it makes a nice starting
spk_0 position to make forward models and predictions of the planets that might result. But I would
spk_0 say that the idea of a magma ocean is quite old. Lidenits who we like to think of as
spk_0 one of the fractious co-founders of calculus started his career as a mining engineer in
spk_0 the Hartz Mountain Silver Mines in Germany. And his job was to keep the mines clear of
spk_0 water so that the mining could go on. But being an inquiring mind, he noticed that the
spk_0 water he was pumping out was quite hot and that the farther down he went into the mine, the
spk_0 hotter it got. And so he made a simple first-order inference, which was the earth is a sphere.
spk_0 It's cold on the outside. It's hotter in the middle. Therefore, through time it's been
spk_0 cooling through its surface boundary condition. Therefore, if you go back in time, it was hotter
spk_0 yet. And he posited that the earth had actually condensed from some kind of a vapor or a plasma
spk_0 and then gradually cooled so that its interior was hotter than its outside. And in fact, it
spk_0 was his work and this wonderful book, Protogaya, which you can still buy and read if you've
spk_0 not, that was later read by Fourier, which inspired him to produce the heat transfer equations.
spk_0 And so I'm very fond of saying that earth science is at the very basis of mathematics. It's
spk_0 not usually a popular thing to say, but I say it anyway. So there we are. So even before
spk_0 we really knew what an atom was, we had the idea of a magma ocean from a different path,
spk_0 I suppose. So are you asking yourself, hopefully I've taken you to this point and you're convinced
spk_0 that planetesamoles might have some water, embryos might have some water, we have some chance
spk_0 of bringing water to the planet. But when we look at the energy that's involved in a giant
spk_0 impact and the fact that the entire planet is melted, how can you possibly retain volatiles
spk_0 through that process? And there are, of course, people who still say that you cannot, but
spk_0 I think that the evidence is that you can. And I'll give you some actual observational evidence
spk_0 before I argue with you on a theoretical basis. So here's some fabulous data from the messenger
spk_0 of mission to Mercury. Now Mercury, before we get into the specifics of this graph, Mercury
spk_0 is a planet, as I'm sure that you know, that has a very large metallic core in comparison
spk_0 to its thin, silicate mantle compared to the other planets. And the going model for that
spk_0 is that Mercury had a very large, creationary impact late in its formation. Actually, it wasn't
spk_0 a creationary, I don't know what the right word is, it was destructive. And it blew off a
spk_0 lot of mercury silicate mantle that was then lost into the sun, leaving a very thin mantle
spk_0 behind. So if that's so, then Mercury is surely the poster child for the giant impact.
spk_0 But even if it's not so, Mercury itself was also built by giant impacts because that's
spk_0 inevitably how planets are built. And there's the one real problem I have with that particular
spk_0 hypothesis for Mercury is that you have to remove the silicate mantle from Mercury's orbit
spk_0 immediately. Otherwise Mercury will sweep off its silicate mantle again as it continues
spk_0 around. So that I think is an unsolved problem, which perhaps you could go solve would be
spk_0 good. So this is a compilation of a lot of different data. And so why are we looking
spk_0 at potassium on the vertical axis and thorium on the horizontal axis? Potassium and thorium
spk_0 are two things that the Messantra mission could measure with its instruments. And their
spk_0 importance is that potassium is volatile. So it's going to act like water. It would like
spk_0 to go into a gas phase sooner than its neighbors. And thorium is not volatile. And so it's going
spk_0 to act more like something that wants to stay on the planet. So the very low slope line
spk_0 with the little blue triangles, that's all lunar material. Lunar material is notoriously
spk_0 dry. It was thought for a long time that the moon was absolutely perfectly dry on an
spk_0 almost part per billion basis. And you can see it in this slope of this line that for a
spk_0 given thorium, not volatile content, there's relatively little potassium or volatile. So
spk_0 it's low volatile content. And then on the much steeper line, you'll see a tremendous
spk_0 amount of data from Mars in black. Down at the very bottom if you can see it, there are
spk_0 in the earth line on the same line. They have about the same amount of volatile elements
spk_0 for non-volatile elements. And the surprise of Messinger, a messenger data relies on that
spk_0 little red circle. Mercury is about as volatile rich as Mars in the earth, which is kind
spk_0 of amazing. So mercury held onto its volatiles despite its most violent birth and its position
spk_0 very close to the Sun. And so that's point number one. There's observational data from
spk_0 Messinger that planets that have been through giant impacts hold onto their volatiles.
spk_0 Now in 2008, was it Alberto Saul at Brown University and Eric Hauri in my department at
spk_0 Carnegie and some other people started measuring materials from the moon and looking for water.
spk_0 And I will tell you, being friends with these people and in the community, people thought
spk_0 it was kind of a silly thing to do. But then they found water and then they became very
spk_0 famous. And so if they were not already, it was a huge fine. They measured little volcanic
spk_0 glass beads, which I've also measured during my PhD. I remeasured a whole bunch of them
spk_0 from Apollo 15. But I never thought to measure them for water. Isn't that a pity? Now
spk_0 I look back on it. And they found water in these little volcanic glass beads. So it was
spk_0 melting inside of the moon and fire-fountaining eruptions that threw magma onto the surface
spk_0 which froze in tiny beads and it had water in it. And they've since gone back and measured
spk_0 additional volcanic glass beads. Now why would the moon be dry? Everyone thought it was perfectly
spk_0 dry. There's metal in the samples that are brought back by Apollo, so there can have been
spk_0 high oxygen activity would have oxidized all the metal. And the moon is thought to be the
spk_0 detritus through into the largest giant impacts. And so it should have been dried, right? Turns
spk_0 out it wasn't dried. Isn't that amazing? So here's data from their latest paper, 2011.
spk_0 And these are three graphs with fluorine sulfur and chlorine, which are also volatile elements.
spk_0 And on the horizontal axis water. And I don't know if you can read it, but it goes from 0,500 to
spk_0 1,500 parts per million of water. And in the big blue blob is mid ocean ridge basalt from the earth.
spk_0 And so what they're showing here is that the orange dots of the lunar glasses plot more or less
spk_0 in the mid ocean basalt field. So they're about as wet as mid ocean ridge basalt's from the earth.
spk_0 Now I hope that there are not very many people who think that the whole moon is that wet. It's
spk_0 thermodynamically, it does not make sense. You can't make some of the rocks that you see.
spk_0 It's that wet all the way through, but there's at least a part of the inside of the moon that's
spk_0 that wet. All right, so here's my little summary slide of wetness on the moon, which is a variation
spk_0 of one that I made for physics today a couple of years ago. So on the vertical axis is water or
spk_0 hydroxyl, which is OH in parts per million by mass, parts per million. And it goes from point
spk_0 001 parts per million up to 100,000 parts per million. Now in the left hand gray box where it says
spk_0 volcanic glass is an appetite mineral grains, these are measurements of volcanic products from the
spk_0 up to about 7,000 parts per million. And then in the next little box, it are from the lunar surface,
spk_0 surface grains that as measured from orbital spacecraft and shadowed crater soils from orbital
spk_0 spacecraft. And I especially wanted to show this here because this when this data was released
spk_0 a few years ago, even my graduate students were confused by the headlines that said the moon has
spk_0 water. And I don't know if any of you were confused or taken aback by those headlines, the moon has
spk_0 water. But my graduate students actually in lab group, they said, well, do they mean like there
spk_0 were rivers like on Mars in the past? No, no, no. And so for comparison next to that over on the
spk_0 right is Sahara Desert sand. And so the surface of the moon is approximately as dry as the Sahara
spk_0 desert. It's pretty dry. But it's just not zero, not zero. So that can't. So when people say damp
spk_0 and wet in geology, it means all kinds of different things, doesn't it? Yeah. And so here at the
spk_0 very bottom, a little picture of the moon down at .001. And that was everyone's favorite bulk moon
spk_0 water content after the Apollo. I shouldn't say everyone. It was a favorite bulk moon content
spk_0 after the Apollo era. And now, how many people think it's more like 100 parts per million in the
spk_0 bulk moon? 100 parts per million is the equivalent of a half of an ocean on the earth. So damp, I don't
spk_0 know what adjective would be best. And then on the right, you see the bulk moon possible range. We
spk_0 don't know how much water is dissolved in the rocks in the mantle. It might be another ocean. It
spk_0 may be five more oceans. Some people think ten more oceans, but I don't know. I think that's a
spk_0 little extreme, but there are arguments in many different directions. But the point of this is that
spk_0 even in the moon, which is supposed to be the driest body round, is damp. And so then the question
spk_0 becomes, now we've built up our earth. And I hope that I've convinced you that it could have retained
spk_0 some water along the way. So did the earliest times of the earth do it look like life magazine said in
spk_0 1952, did it deserve the name Hadeen at that time? I love this because it's got the giant moon
spk_0 looming right over the horizon because the moon was formed on the a couple of earth radii away
spk_0 from the earth. And so it was huge and raised giant tides on the earth and vice versa back then. And
spk_0 or so. And if you have not read it'll oak, halvinos, short story and cosmic comics about when the
spk_0 moon was close to the earth, you should read it right away. It's quite lovely. He talks about how
spk_0 in the old days, when the moon was close to the earth, they would go out in boats on the ocean and
spk_0 they'd set up ladders and they'd climb up and they'd jump and then they would reach the moon and
spk_0 they would mine it for moon milk. And it's just gorgeous, gorgeous story, gorgeous. So the
spk_0 question is, or did it look like this? Did it in fact, did it look like the New York Times said
spk_0 that it did in 2008 and this was based on the Zircon evidence that the oxygen isotopes in these
spk_0 Zircons show evidence for liquid water in the very earliest time of the earth. And I think that this
spk_0 is the answer. I think that very, very rapidly the earth looked like this and it was not violently hot
spk_0 and rapidly convecting. So I just have I think three more slides to take you through the last bit of
spk_0 this talk because I don't like to go over and trying to be careful. So I spend a lot of my time
spk_0 making models of magma oceans and this is a wedge cross section of a hypothetical planet with a
spk_0 core and a magma ocean. And so I set up both the chemical and the physical equations to simulate
spk_0 solidification of the magma ocean. So cooling is happening at the surface boundary layer,
spk_0 cool drips are going downward, they're depositing solids at the bottom. And I,
spk_0 a priori, I decide the sequence of minerals that are going to form. So the bottom is going to be
spk_0 perhaps post-probskyd if it's a deep magma ocean or provskide in Magnesia, Wustite, for those of
spk_0 you who like this glossary of mineral names. And there's some experimental data on how trace
spk_0 elements or bits of water like hydroxyl or carbon might partition into those solidifying minerals
spk_0 and out of the liquid. And so I calculate the equilibrium composition, both composition of the
spk_0 minerals and I partition into them these trace elements. And then as solidification continues,
spk_0 things like water that really don't like to go into these minerals, little bits of them go
spk_0 into the minerals, but more of them stay behind in the evolving liquids. And so as solidification
spk_0 progresses, the evolving liquids become richer and richer in incompatible elements that don't
spk_0 like to be in these minerals, including water. And so water content in the evolving liquids go
spk_0 up and up. And then we allow the water and the carbon dioxide or whatever the phases we're looking
spk_0 at to digass into an atmosphere. And then we track energy transfer out of the atmosphere. And
spk_0 that way we get both a composition and a time scale for solidification of the planet. And so
spk_0 here's the sequence of events in the simplest case. The simplest case sequence is solidification
spk_0 happens from the bottom upward, but slower and slower as solidification continues because
spk_0 there's a thicker and thicker atmosphere which slows down heat transfer to space. So in the
spk_0 beginning heat transfer is very fast, solidification is very fast, but then as gases join the atmosphere
spk_0 it slows down. Now here's a really key step that I haven't mentioned before. The solidification of
spk_0 a magma body like this produces in its later steps heavier minerals than it does in the beginning.
spk_0 And partly that's because all these minerals have got many of them have A site that will take
spk_0 either magnesium or iron. And they all prefer magnesium. And so they build up a magnesium rich
spk_0 layers and iron is enriched in the meltheads that's excluded from the fractionates. But then eventually
spk_0 the minerals are forced to take in more and more iron. So the last ones are quite iron rich and
spk_0 therefore denser than the first ones. They're also a little wetter because they partitioned in a
spk_0 little more water. And maybe there's also more titanium or chromium or other dense minerals. And so
spk_0 so so now what we've done is we've solidified the magma ocean and we've made a pile of a of a
spk_0 solid mantle which is unstable the top of its dense and it's undergoing an undergrow solid state
spk_0 overturn and the dense stuff is going to sink to the bottom and the bottom stuff is going to rise
spk_0 to the top and then the planet will cool down to what I hope will be that ocean on the in the
spk_0 New York Times. And so this is what the models look like. So I apologize for the size of the font.
spk_0 It didn't occur to me to the last minute I had to make it bigger. On the vertical axis is the
spk_0 log of atmospheric water partial pressure. So it's just how much steam is in the atmosphere,
spk_0 what's the pressure of it. And so if you look at zero on that vertical axis to the right is a
spk_0 little red E which is where earth plots right now. If all of its atmosphere was water we're pretending
spk_0 it's all water. And on the horizontal axis is planetary surface temperature. So if you start over
spk_0 at the right at time zero in red or blue two different models you can see that the planetary
spk_0 surface temperature drops very fast. Those curves straight to the left as temperature drops.
spk_0 But then the atmospheric pressure goes up and the rate of cooling goes down and it ends up
spk_0 completely solid at T1 in between 10,000 and a million years. So depending on how much water
spk_0 there is to degas it happens quite fast maybe even just tens of thousands of years to solidify
spk_0 the entire mantle of the earth. But now you recall it's unstable and the dense stuff at the top
spk_0 is going to sink down and the hot stuff at the bottom is going to rise up and it's going to melt
spk_0 a little bit again and you end up up there at T2 on the right at a hot temperature next to two
spk_0 period overturn. And then in these models the surface just cools through its radiate of atmosphere
spk_0 straight to the left. I'm not assuming any atmospheric loss in this. They're very simple models.
spk_0 But in my simple models very simple models and in other people's much more sophisticated models
spk_0 that do the chemistry and the radiate of a convective atmospheric transfer. I've worked with
spk_0 a group at Jet Propulsion Laboratory and with Erykayav and Lamar in Austria. We all seem to agree
spk_0 that we get into the liquid water field order one two three million years. Incredibly fast.
spk_0 Incredibly fast. And so another thing that I like to think about is this cools in all of my
spk_0 models. The pressures end up to be right through the critical point. And so you would end up with
spk_0 this point where you have super critical fluid over the whole surface of the earth as it's
spk_0 collapsing down into an ocean which I think is a really interesting thing to think about. And so
spk_0 here's that same figure again water from zero to 0.7 weight percent in this case and from zero
spk_0 to 0.4 carbon. And the same library of meteorites with both the differentiated and undifferentiated
spk_0 ones. And so the thing that's going to happen to that atmosphere as it passes through the critical
spk_0 point and collapses is it's going to make a water ocean. And so it turns out that if you start with
spk_0 0.1 weight percent of water, a thousand parts per million, that's the equivalent of making a 5 kilometer
spk_0 deep water ocean over the entire surface of the earth. But if you start with this little is 100
spk_0 parts per million, you still get hundreds of meters of deep ocean. And so what we think happened
spk_0 is that the earth probably went through several giant accretionary impacts to reach its current size
spk_0 with magma oceans of varying depths and water contents of varying amounts that would within
spk_0 two or three million years cool all the way down, collapse into a water ocean over the whole
spk_0 surface of the planet. And then along would come the next accretionary impact. So not only do I
spk_0 think that there were oceans on earth very early, but I think there might have been several oceans
spk_0 on earth very early. All right so very briefly a little advertisement for Simone Marci and Bill
spk_0 Bocchi and I are working on models for that tail of accretion. So how much crust can survive
spk_0 through these different ages and how much of the earth's surface is covered by giant impacts.
spk_0 I think it's extremely interesting work that will bear on what happens to the early atmosphere.
spk_0 And this is really just an advertisement friend who's coming to the AGU meeting in December
spk_0 and would like to talk more about what might happen with a series of impacts. And so then here's
spk_0 my final slide where I just want to sum everything up. And on the vertical axis, on the horizontal
spk_0 axis, there's no vertical axis really. On the horizontal axis is billions of years before the present.
spk_0 And I'm sure you can now recite with me the age of that red line 4.568 billion years ago. That's
spk_0 our CAIs, our calcium aluminum occlusions. At the very top in purple, those are some ages that we
spk_0 know. These are actually radiogenic dated ages for core formation in planetesimals for the formation
spk_0 of some of these differentiated meteorites. Planetesimals form very very early. And then models
spk_0 show that they could still have molten interiors after some tens of millions of years if they have
spk_0 a cold crust against space. So the next thing here in yellow are lunar crustal ages. The moon, as I'm
spk_0 sure that you know, when you look up at the moon, you see craters impact basins that are filled with
spk_0 basaltic rock. So that's the most basalt you can see with the human eye in the solar system.
spk_0 And the white material around it is a flotation crust. This is Pledge of Clay's Feldsbar that
spk_0 floated to the top of the lunar magma ocean and made what were called rock burgs. I can't
spk_0 remember it was by Smith or by Wood in 1970 when they posited the magma ocean on the moon immediately
spk_0 upon return of these white rocks by Apollo. So these are the ages that those flotation crust
spk_0 rocks on the moon. They've spanned a very long time 200 million years. In green are geochemically
spk_0 determined ages of differentiation of the earth's mantle. That is that solidification process
spk_0 of the mantle that created different silicate domains that happened very early. Then on the right
spk_0 is the earth's surface. These zircons that date to as far back as 4.4. The ones that have
spk_0 unequivocal evidence for water is date to about 4.3. So what that gives us at the bottom is this
spk_0 purple line. So this is the time of earth accretion from pebble size CAIs through the asterisk
spk_0 miracle up through planetesimals and embryos to make the earth through multiple giant impacts
spk_0 that I think created multiple oceans and cooling events. And then I think the latest possible
spk_0 moon forming impact might have been about 4.41 to give the earth sufficient time to cool down for
spk_0 a final ocean to make these zircons that are preserved today. And so the reason I like this graph
spk_0 and I've had a couple of versions of it in a couple of papers in just the last couple of years is
spk_0 I love thinking about the evolution of the moon and the evolution of the earth on the same scale
spk_0 because what this means is that you could have been standing in a New York Times version of
spk_0 a watery landscape at say 4.36 if you could breathe the air and not you know all the rest of it that
spk_0 would be unpleasant but you could be standing with your feet in something that looks like an ocean
spk_0 and looking up at the moon and knowing that inside the moon was still an annulus of magma ocean
spk_0 that was yet unsolidified. And the reason that the moon would still be a little molten inside
spk_0 is because it has that flotation crust on the outside and the conduction of heat through that
spk_0 flotation crust is the rate limiting step for heat loss and it hasn't managed to cool all the way
spk_0 and the earth has because it didn't have a flotation crust and in fact we are completely ready
spk_0 for habitability to begin by about 4.3 billion years ago and so this is what I mean when I say
spk_0 the fidelity of the timeline in the very first part of the solar system is getting almost as
spk_0 exciting as the fidelity of the timeline that we have near the present. So thank you very much.