≡ Menu

The Likelihood of Massive Exomoons

Are there large moons — perhaps Earth-sized or even bigger — around gas giant planets in habitable zones somewhere in the Milky Way? It’s a wonderful thought given how it multiplies the opportunities for life to find a foothold even in systems much different from our own. Centauri Dreams regular Andrew Tribick recently passed along a new paper that addresses the question in an interesting way, by modeling moon formation and orbital evolution under widely varying conditions of circumplanetary disk composition and evolution.

We’re entering new terrain from this site’s perspective, because I can’t recall going deeply into circumplanetary disks before, at least not in the exoplanet context. But Marco Cilibrasi (Università di Pisa, Italy) and colleagues take us through the necessary background issues. We have two primary models for giant planet formation inside a protoplanetary disk, one being core accretion, when collision and coagulation occurs among dust particles to build up a planetary embryo. The other is gravitational instability, when a gaseous clump collapses into a planet.

We learn from Cilibrasi et al. that both of these approaches to gas giant formation predict that circumplanetary disks will form around planets in the latter stages of their formation. Here we have to be cautious, because circumplanetary disks do not track precisely with the behavior of their larger cousins around stars. The circumplanetary disk (CPD) receives a constant influx of gas and associated dust from the upper layers of the protoplanetary disk, the result of gas accretion onto the forming gas giant. We can take the moons of Jupiter as our local analog of this process. Indeed, understanding a CPD is crucial to make sense of moon formation.

Image: An artist’s rendition of a sunset view from the perspective of an imagined Earth-like moon orbiting the giant planet, PH2 b (a planet confirmed by the Planet Hunters project). The scene is spectacular, but how likely is it that gas giants would have moons beyond Mars size? The answer to the question awaits further work in exomoon detection, but the Cilibrasi paper sees massive moons as a serious possibility. Credit: H. Giguere, M. Giguere/Yale University.

A few interesting things about the Galilean moons: We know that Io is rocky, while the three outer satellites contain a good deal of water ice. The paper points out that icy satellites can only emerge from a circumplanetary disk with temperatures that have dropped below the water freezing point. But simulations of such CPDs show temperatures as high as several thousand Kelvins. This is an indication that the Galilean moons must have formed very late in the process, when Jupiter had substantially cooled and the circumplanetary disk was beginning to dissipate.

Cilibrasi and colleagues performed hydrodynamic simulations that varied the conditions in the circumstellar as well as circumplanetary disks. 20,000 systems were put into play to study the formation, migration and further accretion of the resulting ‘satellitesimals.’ What happens is striking, for because of the continuous influx of dust from the CPD, high-mass satellites form and migrate into the planet, polluting its atmosphere with metals. From the paper:

Our results show that the moons are forming fast, often within 104 years (20 % of the population), which is mainly due to the short orbital timescales of the circumplanetary disc. Indeed the CPD completes several orders of magnitude more revolutions around the planet than the protoplanetary disc material can do around the star at the location of Jupiter. Due to the short formation time, the satellites can form very late, about 30% after 4 dispersion timescales, i.e. when the disc has ∼ 2% of the initial mass.

During this period of formation, the circumplanetary disk continues to cool off, finally allowing icy moons to form. In fact, the Cilibrasi simulations show that 85 percent of surviving moons contain water ice. The authors found that the satellites lost to the forming planet account for 0.3 Earth masses on average, up to 10 Earth-masses in some cases, thus contributing to the abundance of heavy metals found at Jupiter. But the high mass of the satellites formed by these simulations is an encouraging result.

Even though surviving satellites form late, when the circumplanetary disk has dwindled to about 2 percent of its initial mass, the moons that form out of this late process can be massive, as the paper notes, a distribution “peaking slightly above Galilean masses, up until a few Earth-masses.” This should put the moons around close-orbiting gas giants within range of projects like the Hunt for Exomoons with Kepler. As for exomoons in a habitable zone and the interesting astrobiological implications they offer, we can hope that future breakthroughs in instrumentation will allow us to test these findings.

The paper is Cilibrasi et al., “Satellites Form Fast & Late: a Population Synthesis for the Galilean Moons,” submitted to Monthly Notices of the Royal Astronomical Society (preprint).


Comments on this entry are closed.

  • David H January 29, 2018, 12:39

    Something that surprised me about our solar system is the extent to which our giant planets migrated toward and then away from the sun. I suppose this must mean that Jupiter acquired its satellites once it was at a safe distance from the sun, since they clearly have lots of volatiles that haven’t boiled off. I know that Jupiter isn’t expected to move sunward again, but since it’s the kind of thing that nature allows, it might make recipe for habitability. In general we picture the exoplanet systems we discover as being in their present configurations from formation, but that’s almost certainly false. Planets may have traveled back and forth through the habitable zones around stars in their histories. That process could unlock a great deal of chemical potential energy. For example, imagine Titan migrating sunward. Heating up all those organics would inevitably do something interesting.

    • Robert Walker January 30, 2018, 5:36

      The Jupiter moons will end up in the habitable zone once the sun goes red giant. The main issue is that the ice exposed to the vacuum of space would sublime so quickly all the ice would be gone in decades. However I did a bit of a calculation that if a water vapour atmosphere builds up they can last for millions of years and perhaps a lot longer if there is a scummy layer forms on the water to protect it from evaporation.

      See my answer to a question about future life on Europa on Quora and my Ganymede calculation in my answer to “Do water planets exist?”

      • Michael January 31, 2018, 11:13

        The problem with an organic layer is that it would absorb a lot more radiation driving the temperature upwards encouraging more evaporation. Water on the other hand would evaporate and then freeze in space causing a reflective protective layer lowering the temperature.

        As for the Sun’s final demise it may be a good thing for an intelligent species that have made torus habitats. The material thrown off into space is both huge and rich in elements such as oxygen and hydrogen and could be collected with magnetic fields. The Sun’s death could also aid the distribution of life by propelling organisms into deep space, our life could have come here from such an occurrence.

    • Project Studio January 30, 2018, 9:16

      Its really a mystery to me how gas giants could migrate inwardly and outwardly to such a degree, unless it was through close encounters with other massive bodies in the solar system, in which case its is possible to image that their moons could be cast off as planets and possibly planets could be captured as moons. Now we learn that the solar system seems to be quite atypical in that planetary orbits are ‘spaced out’ much more than we see in comparable exo-planetary systems and the range of planetary masses and diameters is across a much wider distribution than typical. This all suggests that our solar system could have an origin story very different from that of other G-class star systems.

      • Mark Zambelli January 31, 2018, 8:16

        Be careful of selection-bias… we haven’t been doing this anywhere near long enough to get a grip on how common our solar system morphology is yet. Just like in the early days when all we were finding were Hot Jupiters (easiest to detect) people thought our system was really odd yet now we know the Hot Jupiter systems are really quite rare… I suspect in a couple decades we’ll think differently about this too.

      • Michael January 31, 2018, 11:06

        They can be pulled in by the mass of material still flowing inwards to the central star and then outwards as the stars fire up.

  • ole burde January 29, 2018, 13:52

    Habitable exomoons raise several interresting questions …Would the atmoshere of a relatively small exomoon be partially protected by the magnetic field of a giant planet ? Would the gasses lost tend to migrate toward the inner moons ? Tidal lock would not be a problem with a orbital period of a few days …so if atmospheres were stable even mars-size moons could be habitable

    • Brett January 29, 2018, 18:10

      I think it would be a mixed blessing. If it was close enough to the gas giant to be in its magnetosphere, it would get protection from the solar wind and cosmic rays – but it would also need its own magnetosphere to protect itself from the gas giant’s Van Allen belts. Both are possible, as with the case of Ganymede (which has a magnetic field and is sitting in Jupiter’s radiation belts).

    • Ashley Baldwin January 30, 2018, 19:28

      Heller et al have looked at this and basically to be in a parent gas / ice giant magnetosphere any moon would be subject to atleast one and probably both of high radiation and tidal heating . The latter leading to runway greenhouse in any moon large enough to possess a terrestrial like atmosphere in what would still have to be the stellar conservative habitable zone. The general view is that in order to sustain a terrestrial like atmosphere for a meaninful time any moon would have to be about atleast three times the mass of Mars ( so about 0.3 Mearth ) .

      Tidal heating is a real issue with a moon calculated as needing to be orbiting outside of between 5-10 planetary radii to avoid it . So as likely as not well outside the planetary magnetosphere . In order to gain anything like the protection Earth has it would thus need to create it own magnetosphere . Can a moon do this ? Simulations vary as to the relative contributions to this on terrestrial bodies made by direct convection in an outer liquid iron core and that created by planetary rotation . An optimal spin rate of sub two Earth days has been posited , which is likely to be much lower than that of a moon orbiting outside of 5-10 planeatary radii. So that leaves primary core convection alone which is in turn dependent on mantle convection. Our one working example ,Earth, can be explained away by its suitable rotation period . But there are simulations that have shown core convection alone as the driver of a potent magnetosphere . So as things stand a large terrestrial moon could theoretically sit far enough out to avoid catastrophic tidal heating whilst creating a protective magnetosphere- and then sustain it over the necessary geological time scales to give rise to life.

      The trouble with this is , is that all the above also depends on the make up of large moons . Core and mantle . Io is the most “terrestrial” like of Jupiter’s large moons in make up , but also the closest and thus subject to substantial tidal heating and radiation bombardment. None starter . The other Galilean moons are further out , though still too close as well bar perhaps Callisto. However this leads to then having very high ice content with only small , rocky , low iron content cores as a consequence of their core accretion formation – both in the milieu of Jupiter and also as a consequence of being beyond the system “ice line” where that planet formed .

      If the Jupiter mass and much larger planets capable of producing terrestrial size / mass moons describe here as likely as not form as did Jupiter , beyond the ice line , before then migrating inwards to the stellar habitable zone ( and stopping there ) , will they be then able to form the kind rocky , iron-cored bodies required necessary to create the terrestrial like conditions necessary for life ?

      Even if all these potential obstacles are overcome there are two other issues to overcome . Could such a peripitetic planet hold onto such “regular” moons during its turbulent perigrinations ( though it might pick up a large terrestrial “irregular” moon in the process ) Secondly ,being ultra massive , it is going to be subject to extended substantial asteroid and cometary bombardment ,along with its moons, as most spectacularly demonstrated by Shoemaker-Levy .

      I’m an optimist and I hope this work points towards the possibility of large terrestrial and habitable exomoons. Knowing they could form is a good starting point .

      Gaia will play a big role as it is likely to discover 1000 gas and ice giants or more orbiting the kind of close stars ( within 100 light years ) that should ultimately allow the detection of larger exomoons as the precision of RV spectroscopy increases with technological advances . Hunting for Exomoons with Kepler might yet evolve into Hunting for Exomoons with Gaia .

      What the HEK?

  • Jean Schneider January 29, 2018, 14:44

    As too often, some authors redicover old ideas.
    About very massive exommons (and even binary planets),
    they were first proposed by Cabrera & Schneider in 2007:
    Detecting companions to extrasolar planets using mutual events
    Astron. & Astrophys., 464, 1133

  • Andrew LePage January 29, 2018, 14:48

    The prospects for habitable exomoons just keeps getting better and better.


  • Julie Novakova January 29, 2018, 16:11

    It would be nice to see specifically the modeled survival rate of the most massive satellites. Their migration rates would also be interesting, since that’s the core of the mystery of the Saturnian system: If the measured tidal dissipation in Saturn is correct, then the inner icy moons must be very young, otherwise they would have started out well within the planet before migrating outward. Either the measurement is off, or they’re younger, formed likely from the debris from a large satellite destroyed by tidal forces or an impact. But not much about moon system architectures involving very massive moons can be derived from that.

    The habitability prospects look good. I love the Heller et al. paper. Looking into the tidal effects and incident planetary radiation is important for habitable moons. By the way, has anyone looked into induction heating for moon? There is a more general recent paper: Kislyakova et al. 2017, available at https://www.nature.com/articles/s41550-017-0284-0

  • Geoffrey Hillend January 29, 2018, 16:51

    I like the idea of a giant exomoon the size of Earth around a gas giant like on the movie Star Wars, but now I have to think that it is very unlikely. Maybe if there was a migration of a Jupiter sized gas giant it, might capture an Earth sized body, but since moons around a gas giant are thought to have formed from an accretion disk, any Earth sized body due to it’s surface gravity and location in the outer solar system would be able to capture a lot more hydrogen and helium so it would have to become the gas giant. A large body in the inner solar system like a formation of super Earth would do the same thing. Two Earth sized bodies might become two Neptunes or Saturn sized objects as both of their gravity competed to grab the surrounding gas around the stellar accretion disk which seems highly unlikely. The largest body becomes the gas giant and what left over, not much, becomes the Moons.
    Also there is the problem with life dealing with the excess radiation from a large electromagnetic field of a gas giant. It would have to orbit far enough away to not have excess heat from tidal gravitational forces from the larger gas giant. I have to think Earth sized exomoons around gas giants are very rare.

    • Michael Fidler January 29, 2018, 21:59

      But would the heat from the young hot Jupiter burn off the hydrogen and helium? A good place to look for these type of systems would be brown dwarfs, their flares would work to light up any planets/moons around them.

  • Harold Shaw January 29, 2018, 17:34

    For liquid water to be present on the surface of the moon, the planet and its moons would need to migrate into the habitable zone. This would have an effect on the atmosphere of the gas giant. Can we make any predictions about secondary effects on the moon?

  • Brett January 29, 2018, 18:13

    That’s good to know that “Mars-sized” is no longer the suspected upper limit on gas giant moons.

    I do have some concerns about them, though. They’d need to be far enough out so as to not be rendered uninhabitable by tidal heating, and then there’s the issue of impacts. That gas giant as mother planet is going to pull in a lot of comets and asteroids. Most of them will miss or hit the gas giant, but the rate of impact on the Earth-sized moon could be much higher than a lone planet like Earth.

    • Brett January 29, 2018, 18:15

      Actually, I wonder if gas dwarf/earth-sized moon combos are more common. A gas dwarf is small enough that an Earth-sized exo-moon could conceivably form from an impact when it’s still accreting hydrogen, and then it would be too small to keep hydrogen itself.

    • andy February 1, 2018, 14:50

      That depends on whether there’s much in the way of comets and asteroids left in the system. I recall there was at one time evidence for an anticorrelation between the occurrence of massive gas giants and the presence of debris discs, not sure if it still holds up though.

      Still, there’s definitely the potential for the impacts that do occur to be more violent.

  • John walker January 29, 2018, 18:57

    As always when the potential for habitable moons of gas giants is discussed, it’s prudent to consider the effects of the attendent very much higher impact rates on evolutionary potential and the biosphere in general. Io-like suppliers of high energy particles and weak intrinsic magnetic fields due to tidal locking could also potentially endanger the exomoon ecosystem/atmosphere.

    Rene Heller and Ralph Pudritz published an interesting paper on super exomoon formation in 2015. https://arxiv.org/abs/1410.5802

  • Andrew Palfreyman January 29, 2018, 19:53

    I suppose quite a lot depends on the inclination of the orbit wrt the planet’s ecliptic, in the usual case where the moon is tidally locked to its parent planet. For zero inclination, there will be periodic warming and cooling of the moon from the star, whereas ninety degrees inclination would expose the moon to a constant stellar flux on one hemisphere only.

    • Mark Zambelli January 31, 2018, 10:56

      The exomoon at ninety degrees would indeed bathe in constant flux but as the planet orbited the star this would constantly change… for the two times per year the exomoons equator lined up with the star it would enter the planet’s shadow once per moon-orbit and experience a shortish day/night just as at Uranus, only on much shorter timescales as we’re refering to a gas-giant in the habitable zone (assume 1 AU and a solar-twin; then these events would be every six months rather than every forty earth-years for Uranus’ orbit). It also means only creatures on the sub-planet hemisphere would see an eclipse and they’d get this twice a ‘year’, six months apart.

      Those months-long stretches of daylight and then night would only affect the poles though, as all other latitudes would have varying day/night cycles depending on the moons orbital period (and when it is high noon at the north pole, the equator would see an orbit-long evening dusk… during the next orbit though the planet would move along its orbit around the star so the star would start to slowly spiral away from the moons north-polar zenith as it slowly set taking one orbital period (‘day’) to revolve around the sky once and then be a little lower toward the horizon; finally reaching it after a six-month spiral… on the equator though the star would move along the entire horizon and dip below the horizon for longer and longer each ‘day’ giving progressively longer ‘nights’ until the opposite would be true half a planet-year later once the planet was on the other side of its orbit.

      How would this affect habitability I wonder? and think of the views.

  • Michael Fidler January 29, 2018, 22:08

    If three quarters of Jupiter or larger planets would have these large moons surrounding them, how much could this add to the habitual exoplanets/exomoons total percentages? Would this also increase the chances for brown dwarfs having habitual planets, I think there is a paper about brown dwarf planets that would receive enough infrared radiation to heat them enough to be habitable.

  • Rob Flores January 30, 2018, 2:27

    I think the conditions that form large moons are mostly exclusive of
    their being HZ conditions. beyond snow line, tidal locking, radiation belts, heavy vulcanism for rocky moons(their formation is favored close to jupiter and larger gas giants).

    I would give greater chance to Jmass x 10 planet orbiting sun like star at mars orbit to have a moon with chance at reasonable HZ. for such a moon, the tidal locking effects would be minimized due to the primary energy source and infrared heat coming from the very large gas giant which would take a long time to cool the heat inherent in its creation. A modest tug of war between primary and x10 jupiter would serve to keep vulcanism alive and recycle the elements of life on the theoretical moon.
    Still a very infrequent occurence overall

  • ljk January 30, 2018, 10:36

    That’s no moon! Oh, wait…

  • BR Delta January 30, 2018, 10:44

    Would it be possible for satellites that large to form satellites of their own?

    • James Franklin January 30, 2018, 16:45

      Whilst it is not impossible, there are serious issues with the “3 body problem” that seem to preclude Moons with Moons. The issue arises from gravitational interactions with the parent planet so the Moon needs to be far enough from it that it could possess a Moon of its own that would not be subject to the gravitational influence of the parent planet.

      As far as I am aware the overwhelming majority of simulations with current models show that the formation of a Moon around a Moon is highly improbable and even when it does happen it tends to be short lived as the gravitational dance that occurs means this second moon with either crash in the parent planet, crash into the Moon or be ejected from the system entirely.

      • Jens Knudsen January 30, 2018, 19:51

        It should be possible to put some numbers on that one.

      • andy February 1, 2018, 14:45

        What’s the current thinking on the origin of the equatorial ridge on Iapetus? One of the potential theories was that it is the remnant of a destroyed “sub-satellite”. Such objects would likely be restricted to young satellite systems (either around young stars, or in the aftermath of a catastrophic instability that “resets” the satellite system) around planets in wide orbits.

    • Ashley Baldwin January 30, 2018, 19:41

      A variant on that theme might be , in terms of a hab zone gas giant orbiting a red dwarf ( less common the round Sun like stars but certainly still possible atleast for Neptune mass), at the resulting closer orbit ( say a generous 0.2-0.3 AU) could it hold onto a moon . My hunch is that it’s Hill zone would be large enough . If so the moon would be tidally locked to the planet rather than the star . For such a moon its rotation might be around ten days or so which is slow in Earth terms obviously but a good deal better for atmospheric circulation than the circa 80-100 day rotation rate induced in the tidally locked parent planet. It’s orbit could even be inclined sufficiently to help reduce the added difficulty of planetary occultation yet still remain stable over long time periods .

  • Kytshar January 30, 2018, 14:52

    Wouldn’t HD 28185 b be the ideal planet for exomoons? After all, it has a minimum mass of 5.7 Jupiter masses and is located in the habitable zone of a very solar-like star.

  • Michael Fidler January 31, 2018, 1:31

    Something just dawned on me, worlds without end. What if there are so many exterrestrial civilizations out there that none of them can become more then Kardashev scale Type II civilization; — also called a stellar civilization — can harness the total energy of its planet’s parent star (the most popular hypothetical concept being the Dyson sphere — a device which would encompass the entire star and transfer its energy to the planet(s)). Since they would not want wars on an galactic scale, we would be the new kids on the block and best to keep us uniformed as to the rules in the Kingdoms of Heaven.

    • ljk February 1, 2018, 11:16

      Ever see the 1951 science fiction film The Day the Earth Stood Still?


      We shall see if there is a Galactic Country Club and if we are allowed in. The lack of a formal invitation so far may be telling as to humanity’s cosmic status.

  • Geoffrey Hillend February 1, 2018, 17:26

    Titan can keep an atmosphere because tidal heating keeps expelling water ice and methane through cryovolcanoes, but it also is far from the Sun so less solar radiation propells a gas molecule much slower so a planet needs less gravity to keep an atmosphere, a planet with a low escape velocity can keep more atmosphere. If Titan formed closer to the Sun, it would loose a lot of atmospheric pressure and gas since its formation.
    A proto Earth sized body in the outer solar system must become a gas giant since it can hold onto the lighter gases hydrogen and helium better which I thought might result in not much gas and dust being left over for a large Earth sized Moon since during the process of becoming a gas giant they might grab most of the gas and dust leaving only enough left for Titan sized moons. This is only a speculation, but it is a limitation that might factor into making Earth sized exomoons rare, especially if we consider the close distances involved. All Jupiter’s Moons are pretty close within a roughly a million miles from Jupiter. Proto Jupiter’s accretion disk might not have a lot of material at greater distances?

  • Geoffrey Hillend February 1, 2018, 17:28

    I like the Galactic country club idea so that we may not be allowed into it until we get advanced enough and develop the technology which will allow us to have access to it.

    • Michael February 2, 2018, 5:21

      Maybe the invite was nailed to that unpronounceable rock ‘Oumuamua’ :).