Planet Nine: Solar System renegade?

We are discussing the origins of Planet Nine.  In part 1 we explored whether it could have been produced during an orbital instability in the early Solar System.  In part 2 we showed that Planet Nine could have been captured from another star.  Here we will explore another plausible idea that is a little less exotic.

Scenario 3.  Planet Nine is a leftover from planet formation (in the Solar System) that was kicked out as Uranus and Neptune grew big

Let’s discuss Planet Nine’s most probable origins story.  This is the idea discussed by Batygin and Brown in their paper.

Let’s rewind the clock to when the Solar System’s planets were forming in the protoplanetary disk.  Gas giants Jupiter and Saturn formed quickly.  Ice giants Uranus and Neptune grew from a population of large ice-rock bodies that were blocked by Jupiter and Saturn.  Uranus and Neptune each underwent at least one humongous collision with another large body; we know this because their spin axes are tilted with respect to their orbital planets (they have significant obliquities, especially Uranus).

The growth of the ice giants is not perfectly efficient.  About half of the icy building blocks don’t end up in a stable planet but instead get kicked out, usually by Saturn or Jupiter.  Planet Nine could be a large leftover that was kicked out.  After being kicked onto a wide enough orbit, external gravitational kicks from the Sun’s birth cluster could shift Planet Nine onto its current orbit, like we saw above.


Cartoon scenario of Planet Nine’s possible origin as a building block of Uranus and Neptune.  The general setup of this idea is based on this paper.

It is a simple and appealing scenario.  It’s entirely consistent with our current vision of how the ice giants formed.  Sure, there are a few details to be worked out but there are no obvious show stoppers.

This idea does not have any problem with the Oort cloud.  Planet Nine must have reached its current orbit early-on, while the Sun was still in its birth cluster.  The Nice model instability would then have happened later and produced the Oort cloud.  Two separate events at different times and in different environments.

[Paragraph of wild speculation.]  If this is how Planet Nine formed, then it fits into a much larger story that connects with exoplanets, planets orbiting other stars.  The most abundant class of exoplanets are so-called “super-Earths”, planets somewhat larger than Earth on orbits smaller than Mercury.  Super-Earths exist around roughly half of all Sun-like stars.  One idea for the origin of super-Earths is that they formed far from their stars and migrated inward.  In that model, the Solar System’s lack of super-Earths is due to Jupiter, which formed quickly and became a migration barrier.  The building blocks of Uranus and Neptune – including Planet Nine – would have become super-Earths if their migration was not blocked by Jupiter.  So, in this (admittedly speculative) story, Planet Nine could represent one of our Solar System’s failed super-Earths!  [End of wild speculation]

The punchline: perhaps the simplest origins story for Planet Nine is that it was a building block of the ice giants that was scattered out while the Sun was still in its birth cluster.

A quick summary of the three formation scenarios

How can we differentiate between formation scenarios for Planet Nine?  Planet Nine’s mass might hold a clue.  During the ice giants’ growth it is generally the smaller planets that are scattered out.  If Planet Nine formed by this process its mass is likely close to 5 times Earth’s.  Building blocks with that mass are the “sweet spot” for forming Uranus and Neptune, so it seems reasonable for Planet Nine to have that mass.  However, if Planet Nine was captured from another star its mass could be higher or lower. If Planet Nine is discovered and its mass eventually measured, a mass near 5 Earth masses (say, between 3 and 10 Earths) would strongly suggest a Solar System origin.  However, if its mass is much higher – in particular if it is more massive than the ice giants – then it is likely to be a captured extrasolar planet.

Leave a comment

Planet Nine: an intruder among us?

We are discussing the origins of Planet Nine.  In part 1 we explored whether it could have been produced during an orbital instability in the early Solar System.  Here we will explore a much more exotic origin.

Scenario 2.  Planet Nine is a planet that formed around another star and was captured by the Sun

The Sun was born in a cluster with many other stars.  It probably looked something like the Trapezium cluster:


Optical (left) and infrared (right) images of the Trapezium embedded cluster of stars. Credit: Hubble Space Telescope/NASA. 

We don’t know how dense the Sun’s birth cluster was.  It probably contained between a few hundred and a few thousand stars.  The cluster lasted about ten million years before expanding and dissolving, sending each star on its own way into the galaxy.  (The Sun is not in a cluster now).

Wide binary stars – systems in which two stars orbit each other on very wide orbits – are thought to form in the following way. Young stars feel the combined gravity of all the other stars as they orbit within their birth clusters. Stars frequently come close to other stars and give each other gravitational kicks. As the cluster expands and dissipates, two stars that happen to be close to each other can find themselves gravitationally bound on a wide orbit.

Stars can also capture planets during the cluster phase if there are a lot of planets floating around among the stars in these clusters.  And there probably are a lot of free-floating planets wandering among the stars.  A dozen or so free-floating (“rogue”) planets have been directly detected.  We don’t know exactly how abundant free-floating planets are (some estimates are very high) but we know that there are a lot of them.

Where do rogue planets come from?  They probably form like most planets do, in disks of gas and dust orbiting young stars.  Then their planetary systems become unstable.  Like the Nice model instability but much much stronger. Here is the cartoon version:


When the orbits of two (or more) Jupiter-like planets cross, there is a series of strong encounters and one or more planets is thrown out into interstellar space.  Check out this awesome animation by Eric Ford and company:

If an instability happens during the cluster phase, the planets may be launched onto star-like orbits within the cluster.  And for every Jupiter-sized planet that is ejected there are several Neptune-sized planets kicked out. We think that the majority of systems of gas giant planets become unstable, creating an abundance of rogue planets.

In its infancy, the Sun’s birth cluster contained only stars (and brown dwarfs) and leftover gas.  Planetary systems were born in disks around most stars.  Many of the stars formed systems of gas giant planets, and most of these became unstable.  By the time the cluster dissipated it must have been teeming with planets.  Like kids at a summer camp, each planet was born to a different star but sent away from home.

Could the Sun have captured Planet Nine as its birth cluster dissipated?  If so, Planet Nine would have formed around a different star in the cluster.  It would be an extrasolar planet lurking within the Solar System.

The answer depends on the orbital dynamics of this complicated system: planets and stars orbiting each other in a cluster as the cluster expands and dissipates.  Calculations suggest that this is a low probability event.  Only 1-10% of stars like the Sun capture a planet, and there is only a few percent chance of that captured planet being on an orbit like Planet Nine’s.  Most of the time, the orbits of captured planets are about ten times wider than Planet Nine’s.

So capture of a rogue Planet Nine during the dissipation of the Sun’s birth cluster is not impossible, but it’s not very likely.  It is at best a roughly 1% event according to current models.  (Remember, of course, that models are imperfect and can change; still, this idea is not looking good at the moment).

But hope remains: there is another way that planets can be captured in clusters, during stellar fly-bys.  Stars in clusters constantly pass near one another.  Of course, stars everywhere in the galaxy pass by one another, but in clusters they pass closer and more often.

If a star passes close enough to a planet-hosting star, it can “steal” the planet.  This only happens when the passage is close, within 2-3 times the size of the planet’s starting orbit.  The captured planet has a much different orbit around its new stellar host: its orbit is much wider and more elliptical.  Here is a cartoon of how this works:


Illustration of how a star can capture a planet from another star.  The key is for the red star to pass closer than 2-3 times the orbital distance of the planet around the yellow star.

Could Planet Nine have been captured during a fly-by with another star?  For this to have happened, another star must have passed relatively close to the Sun.  However, that star could not have passed so close as to disrupt the orbits of the Solar System’s planets. The limit is roughly 100 Astronomical Units; fly-bys closer than that have a good chance of disrupting the Solar System.  This means that the other star’s planet must have been on a much wider orbit than the Solar System’s planets.

Here is one way this could work.  The Sun passed relatively close (say, 200 AU away) to another planet-bearing star.  That star had a planet on a very wide orbit, 100 or more times larger than Earth’s.  Planets on such wide orbits are generally rare; however, in the aftermath of orbital instabilities planets spend time with very wide orbits before being completely ejected.  We can imagine that the planet’s wide orbit was elliptical because it was in the process of being kicked by another, larger planet into interstellar space.

The fly-by was such that the Sun could capture the planet from the other star, but the other star could not capture the Solar System’s planets.  Here is a cartoon view of how this might have happened:


Illustration of how the Solar System could potentially have captured Planet Nine in a fly-by with another planet-hosting star.  While this scenario is speculative it is plausible. For technical details of how this works, see here.

It’s hard to calculate the probability of this scenario playing out, although the pieces fit relatively nicely.  When planetary systems go unstable, the window during which a planet remains on a wide orbit before being ejected is several million years.  The type of star cluster typically lasts 10 million years.  The typical Sun-like star in a cluster akin to the Sun’s birth cluster has an encounter with another star of a few hundred Astronomical Units. Finally, the probability of capturing a planet on a wide orbit is never 100% because it depends on the planet’s actual position.  It can be as high as 30%.  (I am getting numbers from this paper).

I think it is entirely possible that Planet Nine was captured from another star during a fly-by.  For this to have happened requires specific circumstances but nothing too special.  The Sun had to pass close to a star that happened to have a planet on a wide orbit.  That is probably not an exceptional event because about 20% of Sun-like stars have giant planets, most systems of giant planets go unstable, and most instabilities happen relatively early.  To put numbers on it we need more careful models of this process with particular emphasis on Planet Nine.

To sum up, it’s a real possibility that Planet Nine is an extrasolar planet, born around another star then captured by the Sun. 

In part 3 we will discuss what is currently thought to be the most likely origins story for Planet Nine.  We will also think about how to tell these stories apart.


Planet Nine: kicked out by the moody young Solar System?

It’s official: there might be an extra planet in the Solar System.  It’s called Planet Nine.  Like Hansel, it’s so hot right now.

To find it you must trek past Uranus and Neptune into the dark reaches of the Solar System.  It’s on an orbit hundreds of times larger than Earth’s, that takes 20,000 years to complete a single loop around the Sun.

A quick recap: Planet Nine’s existence has been inferred from the strange distribution of small bodies in the outer reaches of the Solar System (so-called scattered disk objects).  These bodies, whose orbits are detached from the main population of small bodies past Neptune (the Kuiper belt), share a peculiar orbital alignment. This is weird: there is only a tiny (0.007%) statistical chance of those six objects randomly sharing this alignment. This is where Planet Nine comes in.  Konstantin Batygin and Mike Brown at Caltech showed that an extra planet could be pulling the strings, using its gravitation to sculpt the orbits of these objects.  To accomplish this, planet Nine’s orbit must be elliptical and anti-aligned with the orbits of the small bodies.  Planet Nine must also be pretty massive, at least several times more massive as Earth.  Here is what its orbit might look like:


Orbits of the six most distant known Kuiper belt objects (magenta).  Their orbits may be sculpted by Planet Nine (orange).  The other planets are all much closer-in, lost in the Sun’s glare. Credit: Caltech/R. Hurt (IPAC)

Planet Nine also naturally explains a different class of Kuiper belt objects with orbits that are almost perpendicular to the plane of the planets’ orbits.  Plus, it’s an awesome idea.  I’m a big fan.

Of course, Planet Nine hasn’t been found yet.  It’s really really hard to find a small cold planet that far away from the Sun.  And it might not even exist (there are good reasons to be skeptical (see here, here and here) until it shows up on camera)!

In a series of three blog posts I’m going to discuss three scenarios for the origin of Planet Nine.  Two scenarios are based on the idea that Planet Nine formed closer to the Sun and was kicked out onto its current orbit. The other scenario invokes the capture of Planet Nine from interstellar space, implying that Planet Nine is of extrasolar origin.  As you will see, one of the scenarios has a fatal flaw but two remain viable. Here we go.

Scenario 1.  Planet Nine was created during an instability in the giant planets’ orbits. 

The orbits of the Solar System’s planets are extremely well-behaved.  Each planet follows a near-circle around the Sun.  To within a few degrees, each planet’s orbit is also located in the same plane.  It’s all very nice.

But there are subtle hints that things were not always so rosy.  One of the biggest advances in Solar System science in recent times is the realization that the Solar System probably went unstable early in its history.  The instability is thought to have been triggered by a gravitational tug of war between the gas giant planets (Jupiter, Saturn, Uranus and Neptune) and a broad disk of planetary leftovers: essentially comets and other pieces of rock and ice that were not incorporated into the planets as they were forming.

The planets’ orbital instability is called the Nice model (because it was developed in Nice, France).  Below is an animation of a computer simulation of the instability.  The animation only shows the outer Solar System – it does not include the rocky planets. The movie shows how, after a delay lasting millions of years, the planets’ orbits go unstable.  When this happens the planets kick the comet-like leftovers all over the place, and end up in a new, wider configuration.  At the end of the movie the planets’ orbits are very close to their current ones.


Animation of the Nice model instability.  Time zero is the moment of instability.  The green dots represent rocky/icy leftovers of planet formation. The circles are the orbits of the giant planets (including one extra ice giant that is ejected).  Credit: David Nesvorny/SWRI.

At the start of the movie there are five giant planets.  In the actual Solar System there are only four (Jupiter, Saturn, Uranus and Neptune).  When the instability happened, one planet was on a very stretched-out, elliptical orbit.  But then that planet disappeared from the movie.  It’s important to realize that that extra planet may have saved the day.  During the instability, the extra planet was kicked around by Saturn and then Jupiter, and  in doing so it prevented those giant gaseous planets from landing in a configuration that would have destroyed the rocky planets (or at least caused a giant collision between two rocky planets).  So, we are glad that extra planet was (probably) there.

But what happened to the extra planet?  It is generally thought to have been launched into interstellar space, gravitationally tossed from the bosom of the Solar System into the cold blackness of the galaxy.  But what if its orbit was somehow changed and it became Planet Nine?

To understand how this could happen we need to look at how gravitational kicks from a planet (say, Jupiter) change another object’s (say, Planet Nine’s) orbit.  This image shows the idea.


Illustration of how an objects orbit changes as it gets successive gravitational kicks from a massive planet (Jupiter in this example).  The object’s orbit gets bigger and bigger but it keeps coming back to the same place.

Each time Jupiter kicks it, Planet Nine’s orbit gets wider.  But it keeps coming back to the same place – its closest approach to Jupiter – until Jupiter finally kicks it so hard that it never returns.

Planet Nine’s real orbit does not look like any of the yellow curves.  Planet Nine spends most of its time far away from the Sun, and it never crosses the orbits of the other planets.

The story changes when Jupiter is no longer the only thing kicking on Planet Nine.  Imagine Jupiter repeatedly kicked Planet Nine until its orbit was pretty big.  At that point, something farther away gave Planet Nine’s orbit a gravitational kick. That external kick  changes the shape of Planet Nine’s orbit so that it no longer goes close to Jupiter.  In this way, Planet Nine could reach a wide orbit that does not cross Jupiter’s.


How an object’s orbit can be changed when there is an external perturbation.  The external kick can change the shape of the orbit and protect it from coming too close to Jupiter.

Where are the external kicks coming from?  From the Solar System’s surroundings: stars passing by within a few light years, and from the combined gravity of stars and gas in the Sun’s neighborhood.

The Oort cloud – a population of comets surrounding the Solar System – was populated during the Nice model instability.  Remember those planetary leftovers, the green dots in the movie?  They were kicked onto wide orbits by the planets, then external kicks changed the shapes of their orbits and rescued them, just like in the image above.

The Oort cloud is very far away: Oort cloud comets spend most of their time hundreds of thousands of Astronomical Units from the Sun (1 Astronomical Unit is the Earth-Sun distance).  That’s a few light years away from the Sun!


A progressive zoom on the orbits in the Solar System.  Credit: M. Brown/Caltech/R.Hurt/C.Powell

If Planet Nine was captured onto its current orbit during the Nice model instability, that means it has the same orbital history as the Oort cloud comets.  This is important.

The external gravitational kicks that we’ve been discussing can be weak or strong.  At the present they are weak, because the Sun is in a quiet part of the galaxy.  But we know that the Sun formed in a cluster with hundreds to thousands of other stars.  And clusters are busy places.  External gravitational kicks in clusters are strong. The stronger the external kicks, the smaller the final orbit of the comet (or Planet Nine) after it is rescued from getting kicked by Jupiter or the other planets.

Let’s put the pieces together.  If the Nice model instability happened early, while the Sun was still in a cluster, external kicks were strong.  The orbits of Planet Nine and the Oort cloud comets would have been separated from the planets before their orbits got too wide.  Their final orbits would have been a lot like Planet Nine’s current orbit.  Score!

But wait.  If that were the case, then the Oort cloud comets should all have the same kind of orbit as Planet Nine.  And they don’t.  Oort cloud comets’ orbits are hundreds of times wider.

If the Nice model instability happened later, when the Sun was out of its cluster, then external perturbations were weak and the Oort cloud comets would have orbits like their real ones.  Score!

But wait.  Then Planet Nine’s orbit should be in the Oort cloud, not much closer-in.

This scenario cannot match Planet Nine’s orbit and the Oort cloud comets’ orbits at the same time.  It must be wrong (unless we are missing something).  Let’s throw it out and move on to the next one.

In part 2 we will explore whether Planet Nine is an intruder from another star.  Finally, in part 3 we will discuss the most likely origin for Planet Nine.



The colors of other worlds

Carl Sagan famously called Earth the “pale blue dot”. Viewed from a large distance, that is what our complex, vibrant, living planet looks like. In the search for life around other stars, we should be looking for other pale blue dots, right?

Maybe not. There is some reason to think that not all habitable planets will be blue. We might be better off searching for pale red dots or pale green dots or even pale infrared dots. In fact, we already know of one blue planet (called HD 189733 b) that is certainly not livable, as its daytime temperature exceeds 700 degrees Celsius (and it’s a gas giant with no solid surface)!


The colors of planets, as measured by the amount of green vs. blue light reflected. Note the blue color of the exoplanet HD 189733 b. Credit: NASA, ESA, and A. Feild (STScI)

The color of an object represents how much each color of light bounces off of an object. Pop quiz: if a ball absorbs blue and green light but reflects red light and I shine a white light on it, what color will it appear? The answer is: red. The red light bounces off and reaches our eyes, while the blue and green is absorbed.

Two key factors determine a planet’s color: its source of illumination and what colors of light the planet absorbs. The illumination comes from its star. The peak of the Sun’s emission is in the yellow-green part of the spectrum. More massive stars are brighter and bluer, meaning they emit more blue light, and lower-mass stars are fainter and redder. Puny stars – called M stars or “red dwarfs” – emit very little visible light and instead are brightest at infrared wavelengths of light that our eyes cannot detect. What visible light red dwarfs do emit is very red. These stars are important because they are the most abundant type of stars in the Universe.

A planet’s reflectance is affected by a number of factors. Let’s take stock of the colors of the Solar System’s planets. Mars is the red planet; its color comes its surface rocks, which are heavily oxidized (rusted, essentially). We can see down to its surface because its atmosphere is very thin. Venus is mostly a featureless bright planet; its clouds are so thick and reflective that we cannot directly see its surface. Neptune and Uranus appear blue because the methane in their upper atmospheres absorb red light but reflect blue. Saturn’s moon Titan’s orange color is thought to be due to complex organic molecules that form a layer of haze in its atmosphere.


Earth as a pale blue dot. This image shows Earth as viewed from the Voyager 1 spacecraft when it was beyond the orbit of Neptune. Credit: NASA/JPL/Voyager (details of the how the image was taken here).

It turns out surfaces of other life-bearing worlds need not resemble the Earth. In fact, the Earth itself did not always look like a pale blue dot. As the Earth’s atmosphere and surface evolved, so too did its appearance. A study from 2007 used models of Earth’s evolution to show just how different Earth used to look. The most important changes in Earth’s appearance came from changes in the composition of Earth’s atmosphere, which was dominated by carbon dioxide until roughly two billion years ago.

Life on Earth is mainly green. Leaves and grass are the most visible forms of life, and their color comes from a single molecule: chlorophyll, the driver of photosynthesis. Photosynthesis is Earth’s primary source of organic matter. Yet photosynthetic plants need not appear green. A different study from 2007 used a broad range of photosynthetic organisms to understand the fundamental rules that govern how photosynthesis is adapted to take advantage of the energy of the parent star. They found that there is a systematic trend: photosynthetic plants around different types of stars should have different colors. Plant life on planets orbiting stars roughly similar to the Sun (so-called FGK stars) should be similar. But given their very low visible light emissions, red dwarfs systematically push photosynthesis on habitable planets to use the infrared and create plants that appear reddish instead of green.


What a familiar landscape might look like if Earth orbited a red dwarf star rather than the Sun. Credit: Tim Pyle, Caltech.

On a more general note, a recent study created a catalogue of the colors of 137 microorganisms that contain pigments. These included photosynthesizers as well as organisms from extreme environments such as deep-sea vents that do not depend on the Sun. There were some common characteristics (e.g., the importance of specific wavelengths associated with water absorption) but their colors were all over the map. This means that planets whose surfaces are dominated by different types of organisms may have very different colors.

It is the oceans that make Earth a pale blue dot. Yet there are other key surface constituents that are imprinted on Earth’s spectrum; desert sands, conifer forests, snow and ice are among the most important. A planet’s color is a mix of all of these and more, and clouds can hide some or all of these features. It is probably overly optimistic to expect that we could infer the presence or absence of life from a distant planet’s color.

And we should beware of over-interpreting other pale blue dots. Are they blue because, like Earth, they have global water oceans? Or, like Neptune, do they have thick methane atmospheres? Or, like HD 189733 b, do they have high clouds containing small glass-like silicate particles? The context – for example, a planet’s atmospheric chemistry depends on how much energy it receives from its star, which can be estimated from its orbit – can answer some of these questions. Yet given the zoo of possibilities, we should not rush to judgment.

The search for other Earths continues. Should we be looking for other pale blue dots, or green or red dots? Given the enormity of the task of finding and measuring the colors of other worlds, maybe we should just be happy with any dots that we get.


Real-life Sci-Fi world 8: the free-floating Earth

Welcome to Real-life Sci-fi worlds.  I use science to explore life-bearing worlds that are the settings for science fiction stories.

This post is a pared-down summary of an article I wrote for Aeon in April 2015 — see the original article here.

The Sun is pretty key for us here on Earth.  The number one goal of exoplanet-hunters is to find planets like Earth orbiting stars like the Sun.  And we’re getting closer: planets have been discovered with orbits at similar temperatures to Earth but around much fainter stars than the Sun.  There is an interesting alternative: worlds with no Sun.  Free-floating planets that roams among the stars.  Gravitationally bound to the Galaxy but not to any star.  Rogue planets.

A handful of free-floating gas giant planets have been discovered.  A recent study found that Saturn- to Jupiter-mass free-floating planets are extremely abundant, and may be more common than stars in the Galaxy!  While that study is somewhat controversial, it is clear that free-floaters are very common.  There are good reasons to think that rogue Earths are probably even more abundant than rogue giant planets.  So the question becomes: what would a free-floating Earth look like?  And could it host life?


What you might imagine that a free floating planet would look like. Spoiler: it doesn’t look like this!

The answer is: yes, rogue Earths could conceivably host life.  The key difficulty is that the planet must maintain a warm enough temperature for water to be liquid somewhere on the planet.  The only available heat source comes from the planet’s internal heat from the decay of radioactive elements such as uranium.  What the planet needs is to hold on to its heat as efficiently as possible.

There are two solutions. First, the planet could have a thick layer of ice covering an ocean.  If the ice is at least about 10 km thick, then the planet can maintain a global subsurface ocean of liquid water.  This is just like Jupiter’s moons Europa and Ganymede, as well as several other minor planets in the Solar System.  I call these icy rogue planets.


An icy rogue planet. Here, Jupiter’s moon Europa is standing in for the icy rogue. Another spoiler: it wouldn’t look like this!

Second, the planet could have a thick hydrogen atmosphere. Hydrogen is a very efficient thermal blanket.  Plus, it does not condense but rather remains in gaseous form even at ridiculously low temperatures.  A free-floating Earth with a thick hydrogen atmosphere can keep its surface temperature above the freezing temperature of water.  The planet could have lakes and oceans (and possibly life) on its surface.  The atmosphere must be thick: at least ten times thicker than Earth’s and probably more like a hundred times thicker.  I call these rogue blanketed planets.


How you might imagine a rogue blanketed planet would look. Here, Neptune is standing in for the rogue blanketed planet. Spoiler: once again, you would be wrong.

So: what would a free-floating Earth actually look like?  The answer is simple: a shadow.  Both types of planets — icy rogue planets and rogue blanketed planets — are exceedingly cold in their outer regions.  They emit virtually no visible light.  They would simply look like a black circle blocking the distant stars.


What a free-floating Earth would actually look like. The upper atmosphere (or ice layer for that matter) is so cold that it does not emit any visible light.

Let’s get to the money question: what kind of life could exist on these planets?  Almost all life on Earth’s surface relies on the Sun.  Specifically, on photosynthetic organisms that use the Sun to generate organic matter (aka food).  But there are organisms on the bottom of the ocean — at hydrothermal vents — that produce organic matter by taking advantage of chemical gradients (differences in chemical potential) that are themselves generated by Earth’s internal heat.  These creatures are called chemoautotrophs.

On rogue Earths, chemoautotrophs could produce the food for more complex organisms.  However, chemoautotrophs are not particularly efficient: they only harvest available energy about one one-thousandth as efficiently as photosynthetic organisms.  So you would need a lot of them to produce food for a whole planet.  Still, hydrothermal vents host entire food chains including weird creatures like giant worms, snails, and strange eels.


The biological community surrounding a hydrothermal vent along the Juan de Fuca Ridge. Credit: University of Victoria.

Free-floating Earths are in principle capable of hosting life, although it may be a while before we can test this idea.  But who knows how close a rogue Earth is lurking to the Solar System.  These planets are so faint that they could well be among our closest Galactic neighbors.

Again, here is the link to the original Aeon article that is a bit more poetic and covers more ground.


Is the Solar System special?

There are at least three aspects of the Solar System that are weird or at least unusual:

  1. No hot super-Earths.  About half of stars contain a planet larger than Earth interior to Venus’ orbit.  We don’t have one.  That puts us in the minority.
  2. Jupiter.  Only about 15% of Sun-lilke stars appear to have gas giants like Jupiter, and most of these are on very stretched out, “eccentric” orbits.  This makes us unusual at the ~5% level.
  3. Life.  Earth is the only planet in the whole freaking Universe that we are sure harbors life.  That is pretty awesome, and it makes us extra-special.

Is there a connection between these things?  I think so!  I just wrote a blog post about it here.  Let me summarize the main ideas.

The key question is: how do “hot super-Earths” form?  I think that these planets — or maybe just their building blocks — formed farther away from their stars and then were driven inward by the gaseous protoplanetary disk.  That process is called orbital migration.  It looks something like this:


If one of those migrating planets gets big enough, it can gravitationally attract gas from the disk and become a gas giant planet like Jupiter.  Gas giant planets are so massive (Jupiter is more than 300 times more massive than Earth) that they carve ring-shaped gaps in the disk.  Any more distant super-Earths that try to migrate inward are blocked by the giant planet:


What this means is that we owe a lot to Jupiter!  With no Jupiter, Uranus, Neptune and Saturn’s core may have migrated inward and invaded the inner Solar System!  Jupiter is our protector!

A nice feature of this model is that it is testable.  We expect an anti-correlation between the presence of Jupiters and hot super-Earths around other stars.  This anti-correlation shouldn’t be perfect but it should be there.  We’ll find out in a few years!

FYI, here is the more eloquent blog post I wrote for  And here is the technical scientific paper.


Real-life sci-ford world 7: Eyeball planets, both icy and hot!

Welcome to Real-life Sci-fi worlds.  I use science to explore life-bearing worlds that are the settings for science fiction stories.

We have already taken a look at hot Eyeball planets.  Today, I have a post on Nautilus’ blog about Eyeball worlds, both icy and hot.  Check it out here.  The article is a bit shorter and more simplified than some of the other ones, but it explains why the first potentially life-bearing planets we find are likely to be eyeballs.

Comments welcome!

Leave a comment

Real-life sci-fi world #6: Pandora (from the movie Avatar), the habitable moon of a gas giant planet

Welcome to Real-life Sci-fi worlds.  I use science to explore life-bearing worlds that are the settings for science fiction stories.  Up today: can the moon of a gas giant planet — like Pandora from the movie Avatar — really be habitable?

Pandora is one of the coolest-ever settings for a science fiction story.  The life-bearing moon of a gas giant planet.  Orbiting one of the nearest stars to the Sun (Alpha Centauri A).  Bonus: giant floating islands!  In my mind, the setting is what made Avatar such a spectacular movie.

Pandora orbiting its gas giant host.  Pandora is supposedly located in the Alpha Centauri system.  Credit: Avatar/James Cameron.

Pandora orbiting its gas giant host. Pandora is supposedly located in the Alpha Centauri system. Credit: Avatar/James Cameron.

Here is what we know about Pandora:

  • Pandora is the fifth moon of the gas giant Polyphemus (which has 14 habitable moons but no rings)
  • Polyphemus is the second of three gas giants and the fourth planet overall from the star Alpha Centauri A, just 4.4 light years away from the Sun.  Polyphemus is slightly smaller than Jupiter so it is probably roughly Saturn-mass.  Polyphemus actually has two “planetoids” that share the same orbit in its stable Lagrange points (L4 and L5). [Remember, this was a ninja move when we built the ultimate solar system].
  • Pandora has 80% of Earth’s gravity.  It is about 45% as massive and 75% as large as the Earth.
  • Pandora’s atmosphere is about 20% denser than Earth’s.  It contains some familiar compounds — like nitrogen, oxygen, carbon dioxide — as well as some exotic ones like xenon and hydrogen sulfide.  The high carbon dioxide content (about 20%) and hydrogen sulfide make the atmosphere highly poisonous for humans.

Let’s get down to business. We are going to tackle the following questions: Can Alpha Centauri A really host a planetary system like the one in the movie?  Can habitable moons like Pandora really form and survive around gas giant planets?  What would the conditions on Pandora really be like?

I won’t write about the floating islands or the actual lifeforms on Pandora.  I think they are spectacular and best left intact in our imagination.  The goal of this post is just to explore the setting.

Binary stars are a mixed bag (details here).  In some cases they can bad for life by destabilizing the orbits of planets in the habitable zone.  In other cases they are neutral and don’t have much effect.  Let’s take a look at the Alpha Centauri system.  Could Pandora and its host planet Polyphemus really exist in orbit around Alpha Centauri A? 

The Alpha Centauri system contains three stars: Alpha Centauri A and B, and Proxima Centauri.  Alpha Centauri A and B are both similar to the Sun in temperature and brightness.  They follow an elliptical orbit around each other.  Their closest approach is about 8.5 AU (1 Astronomical Unit or AU is the Earth-Sun distance) and their average separation is about 17.5 AU.  It takes about 80 years to complete a full orbit.  Here is what it looks like:

The eccentric orbit of Alpha Centauri A and B, with the orbits of the Solar System's planets for scale.  Credit: PHL@UPR Arecibo.

The eccentric orbit of Alpha Centauri A and B, with the orbits of the Solar System’s planets for scale. Credit: PHL@UPR Arecibo.

Proxima Centauri is a red dwarf star about 15,000 AU away from the others.  It is probably on a long orbit around the two more massive stars.  But given that one orbit would take millions of years, we are not 100% sure.  In any case, Proxima is very faint, at only about 0.2% of the Sun’s brightness.  Its gravity might play a role in this story but its light does not.

There are two key things we need to know:

  1. What is the location of Alpha Centauri A’s habitable zone (the “Goldilocks” range of orbital distances where a planet could potentially have life)? and
  2. Given the stretched-out orbit of the binary stars, could a planet stably orbit in the habitable zone?

Here is one estimate of the size of the habitable zones of Alpha Centauri A and B.  Remember, a planet orbiting too close to its star will get fried and a planet orbiting too far away will freeze over.  The habitable zone (in green) is where it’s at for life (even though a planet also needs to have the right kind of atmosphere and surface etc).

Extent of the (optimistic) habitable zones around Alpha Centauri A and B.  From this paper by Heller & Armstrong (2014).

Extent of the (optimistic) habitable zones around Alpha Centauri A and B. From this paper by Heller & Armstrong (2014).

Alpha Centauri A is brighter than Alpha Centauri B so its habitable zone is more distant.  Alpha Centauri B’s habitable zone actually wobbles a little bit because of the extra light it receives from Alpha Centauri A.  But Alpha Centauri B is too faint to affect Alpha Centauri A’s habitable zone.

Now, would a planet’s orbit be stable in the habitable zone?  At closest approach Alpha Centauri A and B come closer to each other than Saturn and the Sun (see image of orbit above).  These stars exert a strong gravitational pull.  Each star can knock the other’s planets out of their orbits!  Wider orbits are the ones in danger because the other star comes closer.

Studies show that a planet orbiting Alpha Centauri A can be stable out to about 3 AU (about 2.5 AU for Alpha Centauri B).  That’s good news!  The habitable zones of Alpha Centauri A and B are both stable!  [Note: this is for a configuration in which the planet’s orbital plane is the same as the orbital plane of the binary.]

Would the planetary system from Avatar be stable?  I can’t find much information on the configuration of the system, but we can improvise a little.  Pandora orbits the gas giant Polyphemus, which is the middle of three gas giant planets in the system.  So, there must be space for at least one more stable orbit farther out.  Let’s see… the habitable zone starts at about 1 AU and a more distant planet would be stable at least about 1.5 times farther out (see this post for details).  Since orbits are stable out to about 3 AU, this is not a problem. The whole 4-planet system envisioned in the story can be stable.

Wait! Astronomers have been searching for planets in the Alpha Centauri system for years.  Recently, an Earth-sized planet was found in a very tight orbit around Alpha Centauri B.  But they have found no signal from gas giant planets around either star!  In fact, there is so much data of the system that astronomers have ruled out the presence of a planet more massive than 4-8 Earth masses in the habitable zone of Alpha Centauri B (gory details here).

For Alpha Centauri A, there is no planet more massive than 10-12 Earth masses in the habitable zone.  The only possible way that a gas giant like Polyphemus could be hiding around Alpha Centauri A is if its orbit is oriented just right.  To avoid detection the planets’ orbits must be almost perfectly face-on.  Specifically, within about 5 degrees of perfect face-on alignment.  This fortuitous setup is very very unlikely — although not impossible of course.

Sorry, Avatar fans. Pandora is (almost certainly) not there.  Polyphemus does not orbit around Alpha Centauri A (unless its orbit is inclined “just so”).  Big frowny face.

Still, there is good reason to think that Pandora-like moons should exist nearby.  Let’s look at the numbers.

About 10-20% of stars like the Sun have gas giant planets.  A lot of those planets are located in or near the habitable zone.  If just one in every twenty Sun-like stars has a gas giant in the habitable zone, then there should be one within 30 light years of the Sun.  That’s not as close as Alpha Centauri (just 4 light years away) but it’s not too shabby!

Motivation restored.  Frown turned upside down. Let’s keep thinking about Pandora.

Can habitable moons like Pandora really form and survive around gas giant planets? 

Planets form in disks of gas and dust orbiting young stars.  Giant planets are mostly made of gas.  As they grow, a gas giant gathers its own little disk of gas and dust.  Moons form within this disk.

Artist's view of a giant planet growing in a gaseous disk around a young star.  The inset shows the disk of gas and dust around the planet itself -- it is in that disk that moons like Pandora are thought to form.  Credit: P. Marenfeld & NOAO/AURA/NSF.  (see also here)

Artist’s view of a giant planet growing in a gaseous disk around a young star. The inset shows the disk of gas and dust around the planet itself — it is in that disk that moons like Pandora are thought to form. Credit: P. Marenfeld & NOAO/AURA/NSF. (see also here)

The disk around a young gas giant planet basically forms a mini-Solar System!  The giant planet is the star and the moons are the planets. We don’t understand all the details of how moons form (or planets for that matter).  But all four giant planets in the Solar System have large moons.  Jupiter even has four of them (the Galilean satellites).  So we know that moons form efficiently.

This is good news: it makes sense for Pandora to have formed around its gas giant host Polyphemus (or around another giant planet).

But there are threats to life on moons like Pandora.  Many threats are indirectly tied to the gas giant host.

The first danger is for the moon to get too hot and to get fried.  What I mean by fried is that all of the moon’s water is vaporized — not good for life!

Compared with Earth-like planets, moons have two extra sources of energy.  The first is tides.  Tidal forces stretch a moon out and generate heat in the interior.  This can generate massive volcanism like on Io, Jupiter’s innermost big moon.  Io is cool to look at but persistent global volcanism (and the extra heat it brings) — not good for life.  We don’t want Pandora to look like this:

Io, Jupiter's volcanic moon.  The insets zoom in on the huge volcano.  Source: NASA/Jet Propulsion Laboratory

Io, Jupiter’s volcanic moon. The insets zoom in on the huge volcano. Source: NASA/Jet Propulsion Laboratory

The second extra source of energy on moons is energy radiated from their parent gas giants.  As a gas giant planet slowly shrinks under its own weight, it loses energy.  This energy is released as heat (infrared light).  For a moon orbiting very close to the gas giant this can be a big heat source and can push the moon toward getting fried.

Moons are more heated the closer they are to their gas giant host.  This is true for both tides and planetary light.  To be safe a moon’s orbit must be at least a few to ten times larger than the gas giant’s actual size (details here).  For scale, Jupiter’s two innermost moons — Io and Europa — have orbital distances of about 6 and 10 Jupiter.

In short: it’s better for a moon to be farther from the gas giant.


There is another threat to life on a moon.  It comes from the bombardment of charged particles.  These particles come mainly from the star, as part of the stellar “wind”.  These particles can act to erode a planet/moon’s atmosphere, change the atmosphere’s chemistry, and do other things that are generally thought to be bad for life.

Earth is protected by its magnetic field.  It creates a little bubble — called a magnetosphere — that shields the planet from the solar wind.

A representation of Earth's magnetosphere.  The planet's magnetic field deflects particles from the solar wind and keeps them from crashing unimpeded into Earth.  Credit: NASA.

A representation of Earth’s magnetosphere. The planet’s magnetic field deflects particles from the solar wind and keeps them from crashing unimpeded into Earth. Credit: NASA.

Small moons probably cannot maintain their own magnetic fields.  But they can be protected from the stellar wind if they stay within their giant planet host’s magnetosphere.  Small moons need to stay close to someone tough (their gas giant host) to avoid getting beat up by bullies (the solar wind).  The closer the moon’s orbit, the deeper within this protective shield the moon is.

But it can be dangerous to order that deep within a gas giant’s magnetosphere.  Jupiter’s moons are protected from the solar wind but can still be bombarded by energetic particles trapped as radiation belts.

The good news: Pandora is large enough that it can probably generate its own magnetic field.  It doesn’t have to worry about this.  The bad news: while it is claimed on the fan site that Pandora has a whopping 13 other habitable moons, it seems unlikely that they can all remain protected from these dangers.  [In fact, it is improbable to have that many large moons orbiting a gas giant and remaining stable — see here.]

Punchline: it is dangerous for moons to be too close to their gas giant hosts because they can overheat due to both receiving light from the gas giant and from tides.  But being close-in can offer some protection from bombardment by the stellar wind.  Since Pandora probably has its own magnetic field, it’s better for it to be farther from Polyphemus.  

What would conditions on Pandora really be like?

What can we reasonably infer about Pandora using astrophysics?  We don’t know all the details of what Pandora looks like (but see this fan-generated globe:)

Comparison of Earth and Pandora.  Note: Pandora is located in the Alpha Centuari system (not "alpha century").  Credit: Okiir.

Comparison of Earth and Pandora. Note: Pandora is located in the Alpha Centauri system (not “Alpha Century”). Credit: Okiir.

First of all, tides shape Pandora’s orbit around Polyphemus.  Pandora always shows the same face to its giant planet: one side of Pandora always sees Polyphemus looming in the sky and the other side never does.

A moon that is tidally "locked" to its planet.  As the moon orbits the planet, it always shows the same face.  People on the planet can never see the green side of the moon.  From Wikipedia Commons (

A moon that is tidally “locked” to its planet, like Pandora must be around Polyphemus. As the moon orbits the planet, it always shows the same face. People on the planet can never see the green side of the moon. From Wikipedia Commons

Our understanding of tides tells us that Pandora’s spin axis must be perpendicular to its orbit around Polyphemus (its obliquity must be very small).  Pandora’s orbit must be near-circular (but not perfectly circular due to gravitational kicks from the other moons).

How long is Pandora’s day?  This depends on Polyphemus’ mass and on the size of Pandora’s orbit.  Since Pandora’s spin is locked to its orbit around Polyphemus, the time to complete an orbit is the length of its day.

If Polyphemus is as massive as Saturn and Pandora’s orbit is similar to that of Saturn’s large moon Titan, then Pandora’s orbit around Polyphemus would take 15 days.  But if Polyphemus is as massive as Jupiter (or a little more massive) and Pandora’s orbit is closer to that of Jupiter’s moon Europa, then Pandora’s day would be just 2-4 days.  In either case, Pandora’s day is longer than Earth’s, possibly quite a bit longer.

Imagine you are standing on Pandora. You are at the right place such that Polyphemus is straight overhead. Let’s see what happens in the sky over the course of a day.  The Sun rises and sets.  So do the stars.  But Polyphemus stays in the same place.  Polyphemus goes from a tiny crescent (just before and after noon) to a fully illuminated circle (at midnight).  As noon approaches, the Sun inches across the sky toward the crescent Polyphemus.  Then the Sun passes behind Polyphemus.  This when Pandora passes through Polyphemus’ shadow.  It’s the only time all day (at this location) to see the stars without a huge, ultra-bright giant planet getting in the way.  It only lasts a few hours but this lull with no heat from the star affects the planetary climate.

A crescent Io orbiting Jupiter.  Image taken by NASA's New Horizons mission.  Credit: NASA/JHU/APL

A crescent Io orbiting Jupiter. Image taken by NASA’s New Horizons mission. Credit: NASA/JHU/APL

There is another big bright object in Pandora’s sky: Alpha Centauri B.  At its closest approach Alpha Centauri B would appear about 3000 times brighter than the full moon!  At its most distant point it is still about 250 times brighter than the full moon. When Alpha Centauri B is in Pandora’s night sky it would be hard to see any other stars!  Polyphemus’ other moons would also appear very bright in the sky.

When Alpha Centauri B shares Pandora’s daytime sky with Alpha Centauri A, it brightens things up a little. Depending on Polyphemus’ orbit, this can represent an increase of as much as 2% in the energy received by Pandora.  This is not bright enough to directly affect Pandora’s climate.  But since Alpha Centauri A is brighter, a planet in the habitable zone of Alpha Centauri B is directly affected by illumination from Alpha Centauri A.

Pandora’s official fan site lists some of its additional characteristics.  It is claimed that Pandora has an axial tilt of 29 degrees and a stretched-out, eccentric orbit.  We know that tides will have driven Pandora to have a low axial and a nearly-circular orbit around Polyphemus.  For this to work it is Polyphemus that must have an axial tilt of 29 degrees (which is similar to Saturn’s).  Since Pandora must orbit Polyphemus in the plane of the gas giant’s equator, this would give Pandora the same effective axial tilt.  The same goes for the stretched-out orbit.  If Polyphemus’ orbit around Alpha Centauri A is non-circular then so is Pandora’s.  The orbit need only be modestly stretched-out, with an eccentricity of a few percent.  This is completely reasonable for the type of planetary system we are expecting.

So there you have it: Pandora as a habitable world.  Although it is extremely unlikely that Pandora actually exists around Alpha Centauri A, it could be habitable.

A final note: there are candidate habitable moons in our own Solar System.  Jupiter’s large moon Europa may have a global ocean lurking under an icy crust.  Heat generated by tides keeps the ocean from freezing over.  It has been proposed that this might be a good home for life.

An artist's view of Europa's subsurface ocean.  Credit: NASA/JPL.

An artist’s view of Europa’s subsurface ocean. Credit: NASA/JPL.

There are several other small Solar System bodies that are thought to have subsurface oceans.  Like Jupiter’s other big moons Ganymede and Callisto.  And the largest asteroid, Ceres.

It has even been suggested that life may exist on Saturn’s large moon Titan.  This life might use hydrocarbons as a solvent instead of water.  Weird but totally possible!

To conclude: I think that the coolest thing about Pandora is that it opens up a whole new door.  Earth 2.0 need not be a planet.  The same goes for Earth 3.0 and 4.0. They could be moons. How cool is that?

In an upcoming post I will address this question in a more general setting.  Specifically, what conditions are required for a moon to be potentially habitable?


Real-life sci-fi world #5: a Dune planet (Arrakis)

Welcome to Real-life Sci-fi worlds.  I use science to explore life-bearing worlds that are the settings for science fiction stories.  Up today: a desert planet like Arrakis from the classic Dune books (and the movie and miniseries).  A tribute for author Frank Herbert‘s birthday (a couple days late).


Watch out for sandworms!

Dune is one of the all-time classic science fiction stories.  It is set on Arrakis, a desert-covered planet.  Arrakis is the only planet in the Dune Universe with melange (aka “the spice”), a life-extending drug that makes it possible for people to travel between the stars.  It is also infested with gigantic super-dangerous sandworms that live underground and are a prodigious source of nightmares.

This post will be a little different than the early real-life sci-fi worlds.  I’m going to use Arrakis as an example of a desert planet. But we will mix it up.  I will discuss the specifics of Arrakis but also some of the more general characteristics of desert planets.

Here are the questions I’ll address in this post.  What would Arrakis’ orbit be like?  Can desert planets like Arrakis be habitable?  How do desert planets like Arrakis lose their oceans?  How common are real-life desert planets orbiting other stars?  And what is likely in store for Arrakis in the future?

Here we go!


What do we know about Arrakis?  Let’s see…. (full details here or here)

  • It is the third planet orbiting the star Canopus, the second brightest star in the sky.  It has two moons.
  • It is mostly covered in sand dunes (hence the name) with some rocky outcrops.  There is no surface water but there are some canals used for irrigation (called qanats).  Some people (the Fremen) collect water in underground reservoirs with the long-term goal of terraforming Arrakis.
  • There is abundant water in the planet’s interior.
  • Salt flats indicate that Arrakis used to have lakes and oceans.
  • The atmosphere is primarily of nitrogen and oxygen in amounts similar to Earth.  The source of oxygen is sandworm metabolism instead of oxygenic photosynthesis.
  • There is some native vegetation similar to that found in deserts on Earth.  There are some grasses, cactii and bushes, as well as a few small animals.

A more general desert planet is simply one that doesn’t have much surface water.  It can have some ponds or lakes but no oceans.  Its surface is dominated by land.  Its atmosphere is dry — there isn’t enough water to provide much humidity on a global scale.

What is Arrakis’ orbit?

Let’s check out Arrakis’ host star Canopus.  Canopus is an “F-type supergiant” star located about 310 light years from the Sun.  It is about 15,000 times more luminous than the Sun!  To receive the same flux from the star, Earth’s orbit would be 123 times wider and it would take 430 years to complete an orbit!  [In other words, Earth’s “year” would be 430 times longer.]

Arrakis’ orbit must be pretty wide.  It could reasonably be hotter than Earth and on a closer orbit.  But it can’t be so close as to preclude habitability.  If we assume that Arrakis receives the same of energy from the star as Venus (which is pushing it a little), its orbital distance would be about 88 times larger than Earth’s around the Sun (it would be 88 “Astronomical Units”, or AU).  Arrakis would take 263 years to orbit Canopus.  That is about how long it takes Pluto to orbit the Sun.

I couldn’t find information on Arrakis’ orbital distance or eccentricity, its obliquity or spin rate.  Given that seasons are not mentioned, Arrakis’ orbit must be pretty circular and its equator is probably close to aligned with its orbit (in other words, it must have a low obliquity and eccentricity).

Can desert planets like Arrakis really be habitable?

Short answer: yes!

It turns out that desert planets like Arrakis might even be more habitable than planets like Earth.  What I mean is, the habitable zones of desert planets may be wider than for planets with oceans.  This was described in a very cool paper in 2011 (paper here, article here).

Desert planets with little surface water can be habitable closer to their stars than “ocean-dominated planets” like Earth with lots of water.  The hottest places on desert planets cool off more efficiently than on Earth-like planets because the humidity is lower; radiation from hot places is a key way that planets (and giraffes) keep cool.  The low humidity in the atmospheres of desert planets also reduces the greenhouse heating from water vapor. Finally, planets tend to lose their water by hydrogen escape from the upper atmosphere.  Since desert planets’ atmospheres are so dry they lose far less water than Earth-like planets.

Arrakis from the video game Dune 2000.  Copyright Westwood Studies/Electronic Arts.

Arrakis from the video game Dune 2000. Copyright Westwood Studies/Electronic Arts.

Desert planets are also better than Earth-like planets at avoid freezing over. When a planet gets cold, water turns to ice.  Ice is very reflective, so the planet absorbs less energy from the star.  This makes it colder, which makes more water turn to ice.  And so on.  This is called ice-albedo feedback (this was important for the oscillating Earth).  Since they have less water, there is a much less ice on a desert planet’s surface when it gets cold.  This effectively stops the ice-albedo feedback from getting carried away.

To summarize: compared with a planet like Earth, a desert planet like Arrakis can remain habitable both closer-in and farther-out from its star.  What is surprising is that it is water itself that reduces the habitability — the ability of a planet to have liquid water — of Earth-like or ocean-covered planets.  Of course, this is a hot topic so there is plenty of debate (see here for a recent perspective).

How do desert planets like Arrakis lose their oceans?

In the Dune universe Arrakis started off as a planet with oceans.  It was biology itself that dried up Arrakis and turned it into a desert planet.  The entire surface water budget of the planet was basically gulped up by sandtrout, the leathery precursors to sandworms.

The sandtrout … was introduced here from some other place. This was a wet planet then. They proliferated beyond the capability of existing ecosystems to deal with them. Sandtrout encysted the available free water, made this a desert planet … — from the book Children of Dune.

Arrakis’ water hasn’t left the planet, it’s just stuck inside of sandtrout, buried in the sand!  This is a case of a very destructive invasive species!  (Like the snakes that were brought to Guam that killed off most native bird species).

In a more general context, where do desert planets come from?  Maybe they just form without much water.  That can indeed happen.  Planets get their water by collisions of objects that condensed in the colder outer reaches of the system.  Planets that form closer to the star can sometimes just not get any water.  The trick is that to be potentially habitable, desert planets still need some water.  They can’t be bone-dry or there would be no liquid water — the requisite for habitability — anyway!

But can Earth-like planets or even ocean-covered planets turn into desert planets?  Well, it depends how efficiently they lose their water.  Water loss is a complicated process that requires a wet upper atmosphere and a lot of very energetic light from the star (X-rays and ultraviolet).  In simple terms, the planets that can lose their water are relatively close to hot stars.  It’s very tricky to calculate exactly but the punchline is: YES, many planets probably can and do lose their water.  These planets start off as ocean-dominated planets and can transition to being desert planets with wider habitable zones rather than just losing all their water too fast.  So in some situations — like very close to the star — an ocean planet must lose its water and become a desert planet in order to become habitable.

It’s easier to make a desert planet out of a rocky planet with some water (a super-Earth) than a gaseous planet (a mini-Neptune).  It turns out that both kinds of planets are very common.  But only small planets are rocky.  Planets smaller than about one and a half to two times as big as Earth.

Bulk densities (y axis) of small extra-solar planets of different sizes (x axis).  Mars, Earth and Neptune are included for scale.  Credit: Lauren Weiss and Geoff Marcy.

Bulk densities (y axis) of small extra-solar planets of different sizes (x axis). Mars, Earth and Neptune are included for scale. Credit: Lauren Weiss and Geoff Marcy (see this paper).  Discussed in detail here.

Small planets are typically rocky whereas larger ones are gas-dominated.  The small ones are the best candidates for being or becoming desert planets.  I could not find any information on Arrakis’ size or mass but it would make sense for it to be similar to Earth.

How common are habitable desert worlds like Arrakis?

Planets exactly like Arrakis must be very rare because stars like Canopusyellow supergiants — are very rare.  Only the most massive stars ever become supergiant stars.  And the yellow supergiant phase itself is just a few thousand year-long stepping stone toward the longer-lasting red supergiant phase.  In time, Canopus will go supernova (see below)!

In short, since only a few hundred yellow supergiants are known, the odds of finding a true Arrakis analog are slim to none.  Plus, since Arrakis takes so long (200+ years) to orbit Canopus and supergiant stars tend to be very far from the Sun, it would be extra-tricky to find a planet orbiting one.

But generic desert planets should be extremely common.  Their habitable zones are wide so there is a lot of real estate in which to find them.  The only requirement is for the planets to be less than about 1.5 times Earth’s size.  At least about 20% of all stars have planets of that size, and at least 10% of stars probably have planets of that size in a broadly-defined habitable zone.  There are almost certainly Earth-sized planets in the habitable zone in our immediate Galactic neighborhood.

But what fraction of these planets is a desert planet?  We don’t have any concrete way of knowing.  In the Solar System there are 2 planets in a broadly-defined habitable zone: one ocean-dominated planet (Earth) and one desert planet (Mars).  It has also been speculated that Venus was a habitable desert planet as recently as 1 billion years ago.

Optimistic estimate: Let’s assume that half of all small planets are desert planets.  There is likely a desert planet around one of the, say, 20 closest stars to the Sun.

Pessimistic estimate: Let’s say that ocean-dominated planets are much more common than desert planets.  Only 1 in 100 small planets is a desert planet.  We must sample the closest several hundred to 1000 stars to find one.  A good place to look may be in a star’s “Venus zone“.  The closest desert planet must still be located within about 50 light years.


Arrakis’ unpleasant future

Like I mentioned, Canopus is a yellow supergiant star.  It started off as a very bright, blue OB star.  These stars are super bright but burn up extremely quickly.  Within 10-30 million years, Canopus used up all of its hydrogen fuel and expanded into a yellow supergiant. [Nit-picky note: this means that Arrakis itself is younger than ~30 million years so the speculation in the Dune Encyclopedia that it lost its water 50 million years ago is off.]

During this expansion Canopus changed color but remained at about constant brightness.  So in Arrakis’ sky, Canopus became drastically larger and changed colors but the energy from the star remained roughly constant.  [Of course, we can speculate that this may have helped Arrakis lose some of its water if the sandtrout were not efficient enough….]

Within a hundred thousand years or so, Canopus will expand further to become a red supergiant (like Betelgeuse).  10-100 million years after that it willl go supernova then collapse into a neutron star or a black hole.

What will happen to Arrakis when Canopus goes supernova?

Best-case scenario: it is flung into interstellar space and survives as a free-floating planet. Arrakis will lose its source of heat but will be well-served by its underground habitats.

Worst-case scenario: it is completely vaporized by the ridiculously intense emissions from the supernova.  Not a happy ending!

There you have it: Arrakis from Dune as a real-life sci-fi world.  I am of course no Dune expert, so if I missed anything please let me know in the comments.  I also only skimmed the surface of some of these issues, so feel free to ask questions in the comments.

, , ,


Real-life sci-fi world #4: Earth around a brown dwarf

Welcome to Real-life Sci-fi worlds.  We are using science to explore life-bearing worlds that are the settings for science fiction stories.  Up today: an Earth-like planet orbiting a brown dwarf.


Planets have been found orbiting all kinds of stars.  Stars like the Sun.  Stars brighter and fainter than the Sun.  Giant stars.  Planets have even been found around brown dwarfs!  [You can poke around in the known extra-solar planets here or here.]

Brown dwarfs are wannabe-stars.  They are just too small.  They can’t generate the huge internal pressure needed to trigger hydrogen fusion in their cores.  So brown dwarfs don’t generate their own internal energy (they do fuse deuterium, but not for very long and that is a pretty wimpy energy source anyway).

With no internal energy source, brown dwarfs spend their time cooling off.  But it takes a while. Billions of years.

Brown dwarfs are between about 10 and 80 times more massive than Jupiter.  But they are about the same size as Jupiter!  Compared with Jupiter, brown dwarfs are just more “squished” under their own gravity.  They are much denser than Jupiter but not much bigger.

Relative sizes of brown dwarfs compared with the Sun, a red dwarf star, and Jupiter.  Credit: MPIA/V. Joergens.

Relative sizes of brown dwarfs compared with the Sun, a red dwarf star, and Jupiter. Credit: MPIA/V. Joergens.

Is there any chance for life around a brown dwarf?  We are confident that planets form around brown dwarfs.  Young brown dwarfs show all the telltale signs of forming planets.  A couple of planets have  been discovered around brown dwarfs, but those planets are too big and too cold for life.  Still, it’s encouraging.

Brown dwarfs are cooler than the Sun so their habitable zones are located much closer-in.  Typically at just a few percent of the Earth-Sun distance (which is defined as 1 “Astronomical Unit” or AU).

Plus, brown dwarfs spend their time cooling off, contracting and getting fainter.  That means that the habitable zone moves inward.  It looks like this:


How the brown dwarf habitable zone changes in time. The green shaded area bounds the habitable zone. At small orbital distances it is too hot and farther out is too cold. The brown dwarf itself shrinks as it evolves — its size is shown in red.  Any planet that comes closer than the Roche limit is torn apart.  This example is for a brown dwarf 4% as massive as the Sun. Adapted from this paper by Emeline Bolmont and collaborators (including me!).

A planet on a fixed orbit cools off along with the brown dwarf.  This corresponds to a horizontal line in the image above.  A planet with an orbital distance of, say, 0.01 AU starts off too hot.  The habitable zone sweeps inward and after about 100 million years it catches up to the planet.  The planet remains in the habitable zone for almost a billion years until it exits past the cold edge.  The frozen planet then cools off indefinitely along with the brown dwarf.

Any planet’s time in the habitable zone is limited, but many planets can spend upwards of a billion years with the right conditions for life.  Another concern is that any planet that enters the habitable zone spent some time too close to the star, in the Venus (runaway greenhouse) zone.  Could such planets retain their water?  [I think so, contrary to this paper.  It depends on the UV and X-ray emissions of brown dwarfs, which are not well known, but we just submitted a paper showing that most planets will keep their water.]

A planet close to a brown dwarf feels very strong tides (from the brown dwarf, not any kind of moon — large moons are not stable for planets close to brown dwarfs).  Tides affect the planet’s orbit and spin.  The planet’s obliquity is quickly driven to zero, such that the equator is lined up with the orbital plane.  The planet’s orbit becomes perfectly circular.  And the planet tidally “locks” and always shows the same face to the star (like for the hot Eyeball planet).  [In this setting tides also cause planets’ orbits to slowly expand.  It doesn’t really affect the story, but see here if you are interested in how it works.]

What would it be like to live on an Earth-like planet around a brown dwarf?  Let’s assume the planet is located in the habitable zone when the system is a billion years old.  For the example from above, that would put the orbital distance at about 0.005 AU.   If Earth’s orbit was the same size, that would put us on the surface of the Sun!  It is a ridiculously small orbit.

So what would be different?  First, as we already discussed, the brown dwarf would always appear in the same place in the sky.  And it would be HUGE!  Our Sun spans about half a degree in the sky.  This planet’s orbit is 200 times closer than Earth’s to an object about one tenth the size of the Sun.  That makes the brown dwarf appear about 20 times larger than the Sun (about 10 degrees across)!  It is as big as a softball held at arm’s reach!  And that “softball” is just hanging there in the sky all the time, never moving…. [cue creepy music]

Earth’s sky is blue because the atmosphere scatters blue light more strongly than red light (this is called Rayleigh scattering).  A brown dwarf emits no blue light.  It barely emits any visible light at all!  Its energy is mainly radiated at infrared wavelengths of light.  There is pretty much no scattering of the light from a brown dwarf.  This means that, to our eyes, the atmosphere would be basically transparent.

The brown dwarf would appear reddish-brown (and gigantic) in the sky.  But look just to the side and you can see the stars! It would basically look like nighttime except in the direction of the brown dwarf!  Clouds would simply be black patches blocking the stars.

The view from a planet orbiting a brown dwarf might look something like this (without clouds):

Author's impression of a brown dwarf 55 times more massive than Jupiter.  Credit: NASA/JPL-Caltech

Author’s impression of a brown dwarf 55 times more massive than Jupiter.  This is basically how the sky would look from a planet orbiting a brown dwarf. Credit: NASA/JPL-Caltech.

Of course, this assumes that inhabitants of these had eyes like ours.  More likely, their eyes would be sensitive in the infrared.  Earth’s atmosphere is almost completely opaque in the infrared except for a few transparent “windows“.  I can imagine these aliens’ eyes being sensitive at several different infrared wavelengths, to gather different types of information.  Maybe these wavelengths would appear as different “colors”.  OK, I’ll stop speculating (for now).


The surface of a habitable planet orbiting a brown dwarf is always illuminated the same.  So the climate would probably be very consistent.  This is not quite as simple as it seems — the planet is also spinning!  To always keep the same face pointed toward the star the planet needs to spin once per orbit.  And a planet in the habitable zone orbits its brown dwarf host in as little as 8 hours!  On average it takes more like a day.  So even though the planet is always facing its star, it spins at about the same rate as Earth!  No one has modeled the climate of this kind of planet so we don’t know exactly what it would be like.

The entire globe could be habitable.  With an atmosphere like Earth’s it would probably be cold on the night side and warm on the day side.  But if the planet’s atmosphere were thicker than Earth’s, then the temperature would be relatively constant everywhere.  Likewise, if the planet’s atmosphere is very thin then the difference in temperature between the day- and night sides would be much larger.


There are a couple of planets known to orbit brown dwarfs.  But they are Jupiter-like planets, not what we’re looking for.  There are all sorts of indirect hints that planets form readily around brown dwarfs.  But we haven’t found them yet.  The main reason is because brown dwarfs are so much fainter than stars, so it’s hard to use them as indirect

There is an interesting strategy in searching for habitable planets orbiting brown dwarfs.  It is based on the transit method.  It works like this.

Take a telescope and stare at some stars (or brown dwarfs).  Carefully measure their brightness as often as you can.  The signal you are looking for is a little dip in the star’s brightness.  The dip repeats once every orbit.  In the best-case scenario it looks something like this:

The dip in a star's brightness when a planet transits its star.   Credit H. Deeg

The dip in a star’s brightness when a planet transits its star. Credit H. Deeg

It’s hard to find Earth by the transit method: the dip in brightness is very small, it only repeats itself once a year, and the odds of Earth’s orbit being lined up right are small (about 1 in 200).

It might be easier to find planets in the habitable zone of brown dwarfs than a planet like Earth orbiting a star like the Sun.  Brown dwarfs are smaller than stars so a transiting planet creates a much deeper dip in a brown dwarf’s brightness.  The transit repeats itself every day instead of every year.  And the chances of the planet’s orbit lining up right are about 20 times higher!  The only problem is that, since brown dwarfs are faint, it is hard to measure their brightness accurately.  Still, this is a promising way to find planets around brown dwarfs (see this paper for the gory details).

Some ballpark numbers.  Within 30 light years of the Sun there are more than 400 known stars (details here).  The vast majority are low-mass (M) stars, also called red dwarfs.  20 are Sun-like (G) stars.  At least 48 are brown dwarfs.  I say “at least” because more are being discovered all the time.

Our Sun's closest neighbors.  Three brown dwarfs (including two in orbit around each other) have only just been discovered (with WISE).  Credit: Penn State University.

Our Sun’s closest neighbors. Three brown dwarfs (including two in orbit around each other) have only just been discovered (with WISE). Credit: Penn State University.

What fraction of brown dwarfs have a potentially habitable planet?  We don’t know.  About half of all red dwarf stars have a planet in the habitable zone.  This fraction is at least 5-10% for Sun-like stars.  Let’s be pessimistic and assume the frequency for brown dwarfs is even lower, just 1%.  There should still be an Earth in the habitable zone of a brown dwarf within 40 light years of the Sun.  Not too shabby!


A group of scientists is discussing the chances for life on other planets.  Except, these scientists are not on Earth.  They are on an Earth-like orbiting in a brown dwarf’s habitable zone.  They sit on old wooden chairs in a dusty room tiled with bookshelves.  Out the window is a starry sky and the giant, looming brown dwarf.

The scientists agree: life is unlikely to exist on planets orbiting stars.  Especially bright stars like the Sun.  The habitable zone is so far away that such planets would only feel very weak tides.  These planets could spin any which way!  The illumination on their surfaces would be changing constantly as the planets spun!  The climate could never be stable on such planets!  Not to mention the variations in illumination over the course of the year caused by the planet’s tilted spin axis (obliquity).  Without tides the planets’ orbits could even be eccentric!  No planet could be habitable with even a modestly-eccentric orbit!

Plus, stars emit ultraviolet light!  That should fry any life on the surfaces of planets orbiting stars. [Note: thank you ozone layer!]

Of course, how could a planet ever develop life in the first place without a hot early phase?  The millions of years before the habitable zone swept inward to include the planet — during which the planet was too close to the brown dwarf — must be responsible for the complex chemistry of life.  Without a hot phase, planets in the habitable zones of stars could never have complex life!

The story ends with a spaceship from an advanced alien civilization — that originated on a planet orbiting a star — approaching the system with evil intentions.  Bad news for narrow-minded thinking!

[Note that this list of criteria that appear vital for life is similar to the Rare Earth hypothesis.  There are 10-20 Rare Earth “factors” that are invoked as being necessary for complex life to develop and survive in a planetary system.  In my mind each of these factors is somewhat arbitrary and most are easily disproved. ]

There you have it: Earth around a brown dwarf.  A Real-life Sci-Fi world.  Special thanks to Franck Selsis and Philip von Paris for some of the ideas in this post.

Questions?  Comments?  Words of Wisdom?

, , ,



Get every new post delivered to your Inbox.

Join 618 other followers