≡ Menu

New Planet Around a Red Giant

Red giant stars have always held a fascination for me, doubtless spurred by an early reading of H.G. Wells’ The Time Machine. Who can forget the time traveler’s journey far into the future after his desperate escape from the Morlocks, millions of days passing in seconds as he flees:

So I travelled, stopping ever and again, in great strides of a thousand years or more, drawn on by the mystery of the earth’s fate, watching with a strange fascination the sun grow larger and duller in the westward sky, and the life of the old earth ebb away. At last, more than thirty million years hence, the huge red-hot dome of the sun had come to obscure nearly a tenth part of the darkling heavens.

We can forgive Wells the mistaken timing — thirty million years won’t account for this! — but still revel in the beauty of the concept. How it must have resonated at the end of the 19th Century. Today, red giants seem a bit more familiar as we’ve learned more about how they happen. And we do know that in perhaps two billion years, the Earth will become uninhabitable as our Sun swells toward the red giant status that awaits it in perhaps five billion years. At that point, the inner planets — Earth included — will probably be swallowed into the furnace.

Would that be the end of subsequent life? We’d like to know more, and the discovery of the tenth exoplanet found around a red giant may play a role in helping us answer such questions. For the more we learn about red giant planets, the better we’ll understand what happens to solar systems after the stellar scenario has changed. Habitable zones move outward as stars move to red giant status, and Alex Wolszczan, a key player on the discovery team of the new planet, thinks this can lead to interesting results even in our own distant future:

“In our solar system, places like Europa — a satellite of Jupiter that now is covered by a thick layer of water ice — might warm up enough to support life for more than a billion years or so, over the time when our Sun begins to evolve into a red giant, making life on Earth impossible.”

Perhaps life gets a second chance. But the planet Wolszczan and colleagues have found using data from the Hobby-Eberly Telescope doesn’t seem to be much of a candidate for habitability. It orbits the K0-giant HD 17092, 300 light years from Earth in the constellation Perseus. The star, roughly twice as massive as the Sun and ten times its size, is circled by its planet every 360 days. The discovery paper estimates the planet’s minimum mass at 4.6 times that of Jupiter.

What we need to do, of course, is expand the sample, and on that score, Wolszczan’s team have compiled a catalog of nearly a thousand giant stars that may host planetary systems. Radial-velocity methods take time to accumulate their data, particularly when dealing with red giants whose planets can take years to complete a single orbit. But three years of work are complete on over 300 stars, making it likely that the new planet is just the first of many we’ll be discussing in subsequent days.

All of which is useful as we try to understand planet formation. Note this interesting passage from the discovery paper (internal references deleted for brevity). It looks at the two major planet formation models — core accretion and disk instability — in light of red giant studies:

Current constraints on a stellar mass dependence of the disk mass and the timescale of depletion of its gas and dust components come from studies of disks around young stars. For example, in addition to the previous work, recent Spitzer observations appear to confirm that disks around intermediate and high-mass stars have lifetimes significantly shorter than 5 Myr. These results have direct consequences for the competing theories of giant planet formation, because the core accumulation scenarios require at least a few million years for a core to form, whereas planet formation from a disk instability can be very short. Clearly, the searches for planets around giant stars have a unique capability to provide the statistics which are needed to decisively constrain the efficiency of planet formation as a function of stellar mass and chemical composition.

These studies thus complement work on other stellar classes and begin to give us an idea of how planetary systems change over time. Their glimpse of our own Solar System’s future may not be as exotic as Wells’, but they should help us piece together a few more parts of the puzzle of planet formation. The paper is Niedzielski et al., “A Planetary Mass Companion to the K0 Giant HD 17092,” to be published in the Astrophysical Journal (preprint available).

Comments on this entry are closed.

  • philw August 3, 2007, 11:26

    As the sun ages and grows hotter there is a simple method (besides carbon nanotube sunscreens) to address the issuee. ‘Simply’ divert a modest asteroid into an orbit where it passes ahead of the Earth in encounter after encounter say a hundred+ years apart and you can SLOWLY expand the Earth’s orbit as the sun ages. You have lots of time to do this. Whatever you do, DO NOT miscalculate. Game over!

  • andy August 3, 2007, 12:30

    Of course, even if we move the Earth out of the danger zone, there’s still the small matter that the interior of the planet is cooling. Once it cools past a certain point, the magnetic dynamo will be in trouble, which would cause problems for the atmosphere. Furthermore, if there is no longer enough internal heat, the plate tectonics and volcanism will start to fail, preventing recycling of carbon dioxide back into the atmosphere from carbonate rocks etc., which apart from causing global freezing, would cause quite a lot of problems for plant life.

  • ljk August 3, 2007, 13:34

    A paper a few years back said that Earth may not be consumed
    by Sol when it turns into a red giant, as the decreased mass of
    our transformed star will cause Earth’s orbit to widen.

    Of course this won’t stop our planet from being roasted, but at
    least it won’t be consumed, for whatever good that will do.

    Then again, the Sol system should have been converted into
    a Dyson Shell ages earlier and whoever is running the show
    by then should know how to keep Sol stable much longer
    than it would last naturally.

  • FrankH August 3, 2007, 14:50

    So when the Sun becomes a red giant Europa will be warm enough to support life? Right. Maybe if it also gains enough mass to retain an atmosphere – otherwise all that warm water will disassociate and leave little more than a rocky core behind.

    I think the chances for habitable moons around any gas giant in a star’s HZ are just wishful thinking. Not only would the “moon” have to be at least Venus sized to retain an atmosphere, it’ll be tidally locked, so little or no intrinsic magnetic field. If it’s deep in the gas giant’s field, the charge particle environment will probably make the moon’s surface uninhabitable. and if it’s orbiting outside the more intense parts of the field, it’ll be exposed to its sun’s field. Plus, the “day” will be many Earth days long.

    The whole concept reminds me of people who call Jupiter a “failed star”…

  • ljk August 3, 2007, 15:54

    The Bad Astronomer Phil Plait comments on why this planet’s
    Earthlike orbit isn’t really like Earth’s:


  • andy August 3, 2007, 17:17

    Ganymede and Titan are probably better bets to become habitable worlds (after a fashion) than Europa: they are more massive (all the better for holding on to an atmosphere with), and Ganymede has a magnetosphere (see below). Also they have more water, so would have greater supplies to replenish atmospheric loss.

    FrankH: the case for habitable moons isn’t quite as bad as you make out… the minimum mass for a moon to retain an atmosphere is slightly above the mass of Mars (if you want tectonics for 5 billion years without tidal heating, the limit goes up to about a quarter of an Earth mass), and it is definitely possible for a moon to have its own magnetosphere inside a gas giant’s magnetosphere: Ganymede’s magnetic field has a strength of about 0.75 microtesla at the equator, compared to 30 microtesla away from the magnetic poles for Earth’s field (factor of about 40 difference), and Ganymede has a slow rotation and is only 0.025 Earth masses. How effective this is at shielding the surface of Ganymede against Jovian radiation, I don’t know, but larger moons would presumably be able to generate stronger magnetic fields.

    Given worlds tidally-locked to their stars can apparently maintain habitable conditions, presumably a slowly-rotating world could also maintain habitable conditions, though it would be interesting to see the kind of adaptations photosynthetic organisms would have to cope with the long nights.

  • Adam August 3, 2007, 22:38

    Hi All

    Greg Benford has suggested, on this board even, that large magnetic “paddles” might be used to reach into the Sun’s core and keep it well mixed with the rest – thus extending the Sun’s hydrogen main sequence life by ~ 50 billion years. Once that ends we might be able to guide the Sun’s entrance on to the helium main sequence and squeeze another ~ 20% of life-time out of it. Past that perhaps some tricky supersymmetric physics could cause the Sun to slowly collapse into super-symmetric particles thus forming something like a neutron star. About a trillion years could be squeezed out of it, then we might catalyse proton-decay or reverse baryogenesis and get a few more trillion out of the old girl.

    But, eventually, we’re going to run out of Sun in about 15 trillion years. As another poster has noted the Earth will become uninhabitable long before we reach that point in time. In about 30 billion years the core will freeze and volcanism will come to an end. But if we can kindle reactions in the Sun we might be able to coax some of the Earth’s core to turn into energy too. If we settled for the current heat out-put of about 30 terawatts the Sun will run out long before the Earth does. If we illuminated the whole Earth with as much heat/light as the Sun blesses us with, the Core would last 37 quadrillion years. Gravity will dip to just 62%, but otherwise things would putter along happily long after the stars die.

    What migh

  • Adam August 3, 2007, 22:42

    I was going to say…

    What might Life get up to in 37 quadrillion years?

    …as for photosynthesis and long nights, well the higher the latitude the longer the day (and nights) get with the seasons. In the high Arctic it stays dark for quite some time and more than a few plants and animals struggle through.

  • andy August 4, 2007, 6:16

    Then again, factors other than core freezing may limit the lifetime of an Earthlike world: the continental area increases over time, increasing the amount of chemical weathering and removal of CO2 from the atmosphere. In fact, from this paper it seems the maximum lifetime for an Earthlike planet is about 6.5 billion years thanks to geological evolution, so unless you’ve got some serious tectonic engineering equipment to hand (which brings up the issue of how to subduct a large fraction of the continental area of the planet in as non-destructive a way as possible…), developing technologies to stabilise the star is a bit pointless.

  • david August 4, 2007, 12:20

    If preserving the earth is the goal, and the sun is one day going to gobble it up, then the solution is simple – get rid of the sun.

    Well, we are talking of the realm of the far future here and if magnetic paddles to stir the sun is feasible then why not a process to dismantle the sun. The material can then be stored and used to power energy generators in earth orbit. Only the energy the earth needs is produced. The generators could have two processes, the fusing of hydrogen into helium, and the splitting of helium into hydrogen, so as to get the full energy from the material.

    To take that a step further we can enclose the earth in a material which captures all the energy hitting it. Both from the outside and the inside. Energy from the stars can be used as well as captured heat energy reflected back from the earth.

    If that still isn’t long enough we can use our advanced von neumann technology to dismantle a few billion stars and store that as well. (I wouldn’t suggest storing all that material in one place though)

    Keeping the core of a planet hot would be child’s play to such a civilization. But we better get cracking, we only got a billion years or so to develop such technology.

  • philw August 4, 2007, 13:24

    Andy cites core tectonics and resultant CO2 management as limiting habitable Earths to a roughly 6-7 billion years lifecycle. If this limit be so, the projected long lifetimes of hypothethized Earthlike planets around long lived red dwarfs may not occur. Given the 1st billion years or so of early bombardment in a new solar system complex, ‘animal’ life forms may be rare because planetary habitability lifetimes are limited and do not approach the age of the galaxy.

  • Adam August 4, 2007, 20:59

    Hi All

    I don’t think we know as much about tectonics as we would need to be dogmatic about the life-time afforded an Earth-like planet. Certainly there may be processes that keep planets evolving long after the life-spans we expect from naive modelling – Io wasn’t expected to be volcanic, nor was a intrinsic magnetic field expected around Ganymede, yet both obtained. The Universe may well surprise us in ways we can’t yet imagine.

    That being said something is needed to explain the Fermi Paradox. I personally think it’s highly likely that we are the first – or close to the first – intelligent species in the Galaxy. Colonizing a Galaxy should happen in a *snap* of cosmic time, but it hasn’t.

  • andy August 4, 2007, 21:28

    Definitely there’s going to be a lot of leeway since we don’t have a very large sample size of terrestrial planets (and the sample size of plate tectonics is one). Though I suspect those results are probably correct to within an order of magnitude, so maybe peg the lifetime of an Earth-mass terrestrial planet at 10 billion years. (Probably depends on how much water you have to start with as well, and wouldn’t apply to ocean planets since the ocean and the rock is separated by the high-pressure ice mantle)

    But you aren’t entirely accurate about Io’s volcanism being unexpected… it was predicted (just) before the discovery: the relevant paper was published on May 2nd, while the Voyager 1 encounter occurred on May 5th 1979…

  • Adam August 5, 2007, 7:13

    Hi andy

    Got me there. But my point was no one expected tidal forces to make a moon quite so volcanic – until we actually went and looked. And who knows what else we might find?

    I’m just a bit wary of ‘easy’ Fermi Paradox answers which assume Life in the Universe is some sort of lucky fluke, a multi-billion-to-one chance that only obtains through a big enough sample with a statistical spread big enough to produce an out-lier like us. I think the Paradox is telling us something more interesting than that – we’re in a very unique historical period for the Cosmos, the beginning of a phase transition.


  • andy August 5, 2007, 10:43

    Likewise I don’t think this is necessarily the explanation for the Fermi paradox.

    I’m not sure that the Fermi paradox really tells us all that much though, except that rather naive view that civilisations go through some kind of Singularity and then commence mega-engineering projects on the scale of star systems and send out rapid and obvious von Neumann probes rampaging through the galaxy, is almost certainly wrong. Alternatively, very few civilisations get to that stage of development (e.g. they go extinct, or their technological development plateaus at a sub-Singularity level), or no-one’s managed to get there yet. I’ll put my cards on the table now and say right out that I’m a Singularity-sceptic, it seems like post-Singularity technology is the new magic wand: the ultimate deus ex machina.

    The issue of technological civilisations at our level of development (no chewing up the scenery or interplanetary colonisation) is still very much open, as is the question of how rare even “primitive” life actually is.

  • FrankH August 6, 2007, 11:08

    Andy – I think the requirements for a habitable planet are already pretty narrow. Make the planet a moon around a gas giant and things get a lot tougher; I’m not saying it’s impossible, just unlikely. So whenever I see a gas giant in a star’s HZ, I pretty much write that system off as having an Earth-like planet.

    As for the Fermi Paradox, I like Geoffrey Landis’ explanation:


  • ljk August 8, 2007, 10:34


    Date: Mon, 6 Aug 2007 20:11:08 GMT (291kb)

    Title: Two Jovian-Mass Planets in Earthlike Orbits

    Authors: Sarah E. Robinson (1), Gregory Laughlin (1), Steven S. Vogt (1), Debra
    A. Fischer (2), R. Paul Butler (3), Geoffrey W. Marcy (4), Gregory W. Henry
    (5), Peter Driscoll (2,6), Genya Takeda (7), John A. Johnson (4) ((1)
    UCO/Lick Observatory, (2) San Francisco State University, (3) Carnegie DTM,
    (4) University of California, Berkeley, (5) Tennessee State University, (6)
    Johns Hopkins University, (7) Northwestern University)

    Categories: astro-ph

    Comments: 32 pages, including 11 figures and 5 tables. Accepted by ApJ

    We report the discovery of two new planets: a 1.94 M_Jup planet in a 1.8-year
    orbit of HD 5319, and a 2.51 M_Jup planet in a 1.1-year orbit of HD 75898. The
    measured eccentricities are 0.12 for HD 5319 b and 0.10 for HD 75898 b, and
    Markov Chain Monte Carlo simulations based on derived orbital parameters
    indicate that the radial velocities of both stars are consistent with circular
    planet orbits. With low eccentricity and 1 lessn than a less than 2 AU, our new planets have
    orbits similar to terrestrial planets in the solar system. The radial velocity
    residuals of both stars have significant trends, likely arising from substellar
    or low-mass stellar companions.

    http://arxiv.org/abs/0708.0832 , 291kb

  • ljk September 13, 2007, 8:49

    Earth could survive a red-giant Sun

    In about five billion years the Sun will run out of hydrogen fuel and swell into a red giant star over a thousand times its current volume before shrinking back into a white dwarf. No-one is quite sure whether Earth is close enough to be swallowed up by a bulging Sun or whether it can avoid a fiery death. But now an international team has spotted a planet in a distant solar system that appears to have survived its star’s red-giant phase, even though its original orbit would have been similar to Earth’s (Nature 449 189).

    Full article here:


    And from the NYT:


  • ljk January 23, 2008, 10:11

    Testing planet formation theories with Giant stars

    Authors: Luca Pasquini, M.P. Doellinger, A. Hatzes, J. Setiawan, L. Girardi, L. da Silva, J.R. de Medeiros

    (Submitted on 22 Jan 2008)

    Abstract: Planet searches around evolved giant stars are bringing new insights to planet formation theories by virtue of the broader stellar mass range of the host stars compared to the solar-type stars that have been the subject of most current planet searches programs. These searches among giant stars are producing extremely interesting results. Contrary to main sequence stars planet-hosting giants do not show a tendency of being more metal rich. Even if limited, the statistics also suggest a higher frequency of giant planets (at least 10 %) that are more massive compared to solar-type main sequence stars.

    The interpretation of these results is not straightforward. We propose that the lack of a metallicity-planet connection among giant stars is due to pollution of the star while on the main sequence, followed by dilution during the giant phase. We also suggest that the higher mass and frequency of the planets are due to the higher stellar mass. Even if these results do not favor a specific formation scenario, they suggest that planetary formation might be more complex than what has been proposed so far, perhaps with two mechanisms at work and one or the other dominating according to the stellar mass. We finally stress as the detailed study of the host stars and of the parent sample is essential to derive firm conclusions.

    Comments: IAU 249: Exoplanets: Detection, Formation and Dynamics J.L. Zhou, Y.S. Sun & S. Ferraz-Mello, eds. in press

    Subjects: Astrophysics (astro-ph)

    Cite as: arXiv:0801.3336v1 [astro-ph]

    Submission history

    From: Luca Pasquini [view email]

    [v1] Tue, 22 Jan 2008 10:20:22 GMT (233kb)


  • Robert January 25, 2008, 14:30

    And whos to say that life requires a carbon based set of guidelines and lots of water? Maybe there are replicator proteins sitting around in a pool of chilled gas on Uranus that is just waiting for the right kick to get going. Like the Red Dwarf stage of our sun.

    We are based on a chiral set of structures, and there is a whole other half of potential life variations besides all the possible ones given our classification; those of which we can’t even fathom a percentage of.

    I think that the ability of life to exist is beyond our ability to comprehend completely. It would be naive to think otherwise. But the fact that our current life situation is unique compared to other situations that have potentially, theoretically evolved has been completely unaddressed by this post ^_^.


  • ljk July 3, 2008, 16:37

    Planetary Companions to Evolved Intermediate-Mass Stars: 14 Andromedae, 81 Ceti, 6 Lyncis, and HD 167042

    Authors: Bun’ei Sato, Eri Toyota, Masashi Omiya, Hideyuki Izumiura, Eiji Kambe, Seiji Masuda, Yoichi Takeda, Yoichi Itoh, Hiroyasu Ando, Michitoshi Yoshida, Eiichiro Kokubo, Shigeru Ida

    (Submitted on 2 Jul 2008)

    Abstract: We report on the detection of four extrasolar planets orbiting evolved intermediate-mass stars from a precise Doppler survey of G and K giants at Okayama Astrophysical Observatory. All of the host stars are considered to be formerly early F-type or A-type dwarfs when they were on the main sequence.

    14 And (K0 III) is a clump giant with a mass of 2.2 M_solar and has a planet of minimum mass m_2sin i=4.8 M_Jup in a nearly circular orbit with a 186 day period. This is one of the innermost planets around evolved intermediate-mass stars and such planets have only been discovered in clump giants.

    81 Cet (G5 III) is a clump giant with 2.4 M_solar hosting a planet of m_2sin i=5.3 M_Jup in a 953 day orbit with an eccentricity of e=0.21. 6 Lyn (K0 IV) is a less evolved subgiant with 1.7 M_solar and has a planet of m_2sin i=2.4 M_Jup in a 899 day orbit with e=0.13.

    HD 167042 (K1 IV) is also a less evolved star with 1.5 M_solar hosting a planet of m_2sin i=1.6 M_Jup in a 418 day orbit with e=0.10. This planet was independently announced by Johnson et al. (2008, ApJ, 675, 784). All of the host stars have solar or sub-solar metallicity, which supports the lack of metal-rich tendency in planet-harboring giants in contrast to the case of dwarfs.

    Comments: 23 pages, 6 figures, accepted for publication in PASJ

    Subjects: Astrophysics (astro-ph)

    Cite as: arXiv:0807.0268v1 [astro-ph]

    Submission history

    From: Bunei Sato [view email]

    [v1] Wed, 2 Jul 2008 04:05:29 GMT (111kb)