≡ Menu

A Gaseous, Earth-Mass Transiting Planet

Search for one thing and you may run into something just as interesting in another direction. That has been true in the study of exoplanets for some time now, where surprises are the order of the day. Today David Kipping (Harvard-Smithsonian Center for Astrophysics) addressed the 223rd meeting of the American Astronomical Society in Washington to reveal a planetary discovery made during the course of a hunt for exomoons, satellites of planets around other stars. Kipping’s team has uncovered the first Earth-mass planet that transits its host star.

Just how that happened is a tale in itself. Kipping heads up the Hunt for Exomoons with Kepler project, which mines the Kepler data looking for tiny but characteristic signatures. Transit timing variations are the key here, for a planet with a large moon may show telltale changes in its transits that point to the presence of the orbiting body. In the case of the red dwarf KOI-314, it became clear Kepler was seeing two planets repeatedly transiting the primary. No exomoon here, it turned out, but as David Nesvorny (Southwest Research Institute) puts it:

“By measuring the times at which these transits occurred very carefully, we were able to discover that the two planets are locked in an intricate dance of tiny wobbles giving away their masses.”

The star, located about 200 light years away, was orbited by KOI-314b, about four times as massive as the Earth and circling the star every thirteen days. Transit timing variations, it became clear, flagged the presence not of an exomoon but another planet further out in the system, dubbed KOI-314c. And while the latter turns out to have the same mass as the Earth, it is anything but an ‘Earth-like’ planet. KOI-314c is, in Kipping’s words, ‘the lowest mass planet for which we have a size *and* mass measurement.’ With size and mass in hand we can work out the density. This is a world only thirty percent denser than water, as massive as the Earth but sixty percent bigger, evidently enveloped in a thick atmosphere of hydrogen and helium.


Are we dealing with something like a ‘mini-Neptune,’ perhaps one that has lost some of its atmosphere over time due to radiation from the star? Whatever the answer, KOI-314c forces us to re-examine our assumptions about small planets and how they are made. In a recent email, Kipping told me:

“If you asked an astronomer yesterday what they would guess the composition of a newly discovered Earth-mass planet was, they probably would have said rocky. Today, we know that is not true since now we have an Earth-mass planet with a huge atmosphere sat on top of it. Nature continues to surprise us with the wonderful diversity of planets which can be built.”

Image: Exomoon hunter David Kipping, whose fine-tuning of TTV (transit timing variations) is opening up new possibilities in exoplanet and exomoon detection.

Exactly so, and what a potent lesson in minding our assumptions! Kipping continues:

“We now know that one can’t simply draw a line in the sand at X Earth masses and claim “everything below this mass is rocky”. It hints at the fact that the recipe book for building planets is a lot more complicated than we initially thought and so far we’ve perhaps only been looking at the first page.”

A humbling notion indeed. Finding solar systems like our own, with rocky worlds in an inner region and gas giants further out, has proven surprisingly difficult, making us think that our system may be anything but typical. Now we’re examining a small planet that forces questions about planetary mass and composition and leaves us without easy answers. Its discovery highlights the significance of transit timing variations, a technique that is clearly coming into its own as we reach these levels of precision studying low mass planets. We’re still looking for that first exomoon, but the hunt for these objects is pushing the envelope of detection.

Kipping’s email mentioned yet another significant fact. KOI-314c circles a red dwarf close enough for detailed observations with the upcoming James Webb Space Telescope. The size of the atmosphere around this world could make it ideal for detecting molecules in its atmosphere, adding to our knowledge of this particular planet but also giving us another venue on which to sharpen the tools of atmospheric characterization. A worthy find indeed, and a reminder of how much we have yet to discover in Kepler’s hoard of data.


Comments on this entry are closed.

  • Adam January 6, 2014, 16:39

    The press release at HEK’s website links the discovery paper:


    …for anyone wanting details.

  • Rob Flores January 6, 2014, 17:10

    This certainly a landmark discovery. If more data points come in
    does this validate some ideas of protoplanetary disks. Particularly the idea that with a dwarf central star the lighter elements H2, He, are not totally expelled from regions near these stars, as compared to heavier K and bigger
    stars hotter emission.
    At least it’s a world that is explorable en vivo, and it’s not raining
    granite droplets.
    Speaking of Kepler and other attempts at Exo-Planet search.
    there is a question about the statitical likelyhood co-planarity (if
    that is the term) in exo-systems. That is all the planets of our solar system
    lie within 2 degrees or so from our suns rotational plane. Is there enough
    data in Kepler transits to give an idea (at least for the short period finds).
    of the likelyhood of this 2 degress result among random stars systems.

  • Peter Chapin January 6, 2014, 17:14

    While individual finds like this are fascinating, the real win will come from statistical studies of many systems. Just how unusual is our system and in what ways? That’s a question I’d like to have answered.

  • Lepton January 6, 2014, 19:43

    IMHO, it not surprising that a mini star makes a mini Neptune. H and He are everywhere, with a dim star unable to heat/blow them afar, they will naturally accumulate on any rocky/metal core. What is a big deal about it?

    We are still seriously short on data for long period (1y+) exoplanets around Sun like stars. Without those data, it is hard to say whether our solar system is typical or not.

    It is very unfortunate that 2 gyros are gone on Kepler. :-(

  • NS January 7, 2014, 0:52
  • Michael January 7, 2014, 1:43

    …’And while the latter turns out to have the same mass as the Earth, it is anything but an ‘Earth-like’ planet. KOI-314c is, in Kipping’s words, ‘the lowest mass planet for which we have a size *and* mass measurement.’ With size and mass in hand we can work out the density. This is a world only thirty percent denser than water, as massive as the Earth but sixty percent bigger, evidently enveloped in a thick atmosphere of hydrogen and helium.”

    I am not sure how old the Star is, our early Earth had an H/He atmosphere as well. Perhaps it is not H/He but water/steam, the planet could have come from further out after having accumulated ice and then it started to boil off as it got nearer the Star.

  • Lionel Ward January 7, 2014, 4:10

    An earth mass star with an atmosphere of helium and hydrogen?

    Surely a very surprising find!

    Think about atmospheric escape!

    Earth’s gravity is too low to hold onto hydrogen and helium, so how is it possible for this planet to have a He and H2 atmosphere?

  • Enzo January 7, 2014, 4:42

    This is very interesting and a hint that terrestrial size planets can carry really deep atmospheres, even deeper than Venus.
    I don’t understand the reference to hydrogen and helium : at 1 earth mass, all hydrogen and helium should have escaped especially at the given high temperature.
    See here :

  • David January 7, 2014, 5:12


    Considering that M dwarfs make up ~70 % of all stars and that
    these planetary systems are different from the Sun, we can now say that
    our Solar System is unusual in comparison to all stars. However, what you say is still true with regards as to whether our Solar System is unusual compared to planetary systems found around other Sun-like stars.

  • Harry R Ray January 7, 2014, 10:49

    How would astronomers ELIMINATE the very slim possibility of a 1.6Re solar sail positioned at 0.001AU closer to the star than KOI314c is, whose perpose is to cool the planet from its estimated 220F temperature down to 60F? Can this be done with existing Kepler lightcurve data, Spitzer or Hubble observations, or, Would we have to wait for JWST? This is just an intellectual exercise, since the ability of giant solar sails to cool down planets appears to have already ruled out theoretically!

  • Ronald January 7, 2014, 12:28

    Further to other commenters (partic. Michael, Enzo, Ward): how certain is it that this gaseous envelope is indeen H and He?
    With such low mass/low gravity and a surface temp. of about 100 C, it seems rather unlikely that it would be H/He.
    Could it be water vapor instead?

  • Lepton January 7, 2014, 14:08

    NS, Kepler is one of my favorite mission, high up there with Voyager 2, etc.

    IIUC, K2 mission can’t point to the same star field of K1, so, we need to restart from zero to accumulate data in new fieldS. But, we might get lucky.

    David, you are right. I should have phrased it like “typical or not among planetary systems around Sun like starts”.

  • andy January 7, 2014, 15:24

    We are still seriously short on data for long period (1y+) exoplanets around Sun like stars. Without those data, it is hard to say whether our solar system is typical or not.

    Various RV surveys have now been running long enough to find Jupiter-analogues and they appear to be fairly rare, occurrence rate somewhere in the single-digit % range.

  • Volucris January 7, 2014, 18:40

    Wonderful! A deluge of questions come to mind: is it possible to get a handle on the age of the host star? Does it spin fast? Does it flare? Is it quiet? To which galactic population it belongs to? How metallic is it? Tracking the age of the star could give clues whether the planet is in process of boiling away it’s atmosphere, or if it by some unfathomable feat is actually holding on to it. Also, if the planet is boiling away hydrogen, I’d think the atmosphere was enriched in somewhat heavier helium and deuterium. I wonder if JWST will be able to resolve some information on deuterium (and helium, and heavy water) proportion of the atmosphere. That could give clues on how big the planet was to begin with…

  • Michael January 8, 2014, 3:47

    If we knew the age of the Star it would tend to indicate whether it is H/He or a water/steaming atmosphere. H/He are lost quite quickly due to their volitility. With a water/steam atmosphere pressure can have a huge effect on the size of the atmosphere where as with a H/He one it is not as pronounced.

    Paul, do you have as other commentators have written any further information about the system, orbital distances, star mass, luminosity etc. There is the possibility that the planet has migrated inwards to setup this 5/3 orbital resonnance with the other inner planet. The HBZ is quite narrow in red dwarf star systems, so it may have been in the ice line for a while as it migrated inwards collecting ice which is now boiling away. A look at its spectrum for oxygen and hydrogen could indicate water is the culprit as opposed to H/He gases.

  • henk January 8, 2014, 4:58


    Than we just need to hope that jupiters are not needed for life to exist on inner terrestrial planet.

    It is kind of funny that something that is very common like a mini Neptune are not in the solar system but something like Jupiter is.

  • Paul Gilster January 8, 2014, 11:23

    Michael, re further information, here’s David Kipping’s paper:


    which should supply the information you need. I’d dig it out myself but am currently down for the count with sinus trouble.

  • Michael January 8, 2014, 14:12

    Hi Paul,

    Sorry to hear you are unwell, hope you get better soon and thanks for the paper link.


  • Ashley Baldwin January 8, 2014, 14:54

    If the atmosphere is indeed somehow hydrogen rich, and hydrogen being a great greenhouse gas, we are indeed going to have to review the size of the HBZ around M stars at least.

  • FrankH January 8, 2014, 16:18

    Pineda et al 2013: http://arxiv.org/pdf/1302.6231v2.pdf

    Has the best data for KOI-314 (distance and mass are given within 10%) , but no age estimates.

    IF the planet really has a H/He atmosphere, it’s probably a mini-Neptune that has mostly evaporated and we’re detecting the core. A water/steam atmosphere would not be out of place for an evaporating mini-Neptune, either.

  • andy January 9, 2014, 14:02

    @FrankH: evaporation does seem to be the simplest explanation for the density trends seen in low-mass systems. I strongly suspect that the rocky exoplanets seen so far are chthonian planets rather than genuine terrestrials. In yet another parallel between the exoplanetary systems and the gas giant satellites, that seems to also work as an explanation for why there’s a terrestrial “planet” orbiting Jupiter.

  • coolstar January 9, 2014, 20:05

    Very nice work, but somewhat over stated. From previous work we already knew that mini-Neptunes and super-earths show a diversity of densities and that some of the smallest planets could not be rocky. Indeed, Marcy and collaborators show the second such study at this same meeting, and show a large scatter in the radius vs mass relation for this most common class of planets. J. Swift and his collaborators have also shown that most of these planets around M dwarfs seem to have formed beyond the frost line so this result, while very important, is not that surprising.

  • Ronald January 10, 2014, 6:33

    andy: and I suspect that you are right.
    This graph by G. Laughlin, http://oklo.org/2013/11/16/all-over-the-map/, nicely shows the cluster of super-earths/mini-Neptunes and Neptunes (ice giants) as a continuum, with the bigger ones being less dense, probably because their gaseous envelopes haven’t evaporated (yet).

    So, there could be just 3 main categories left: true terrestrials, midsized (Neptunes and mini-Neptunes with varying degrees of gas envelope erosion), gas giants.
    Which would also imply that the most common type of planet, the super-earth/Mini-Neptune, isn’t really absent from our own system, it is the same as the Neptunes, just more eroded, which hasn’t happened here.

    Leaves the crucial question: when a planet starts its existence as a Neptunian and then becomes a ‘secondary’ terrestrial through erosion, can such a planet then also become a suitable habitable planet (surface water, earthlike primordial atmosphere, plate tectonics, …)?

  • Ashley Baldwin January 10, 2014, 13:27

    Interesting point Ronald. Is it a true secondary terrestrial though? Density only a third greater than water . Does it have the “terrestrial” raw materials to create the silicate dominated mantle necessary to interact with water to produce tectonics and a secondary atmosphere ? If it formed beyond the ice line before later migrating , as seems likely , does it have the iron to form an active core along with the radioactive elements necessary to help keep up its internal temperature and drive those tectonics.

  • Michael January 11, 2014, 4:09

    After a little calculation I find that the planet KOI-314c orbits at around 0.15 au’s from it’s star which puts it about 0.04 au’s past the inner edge of the HBZ. This is a neat program to see the planets position, it also allows you to see the width of the HBZ and the age that these types of star can reach!


    With that in mind it is quite possible to have a water world after having migrated inwards due to orbital resonances to start boiling off, it may also just be a giant water world and not be off gassing due to pressure of an overlying dense N,Co2 atmosphere. Off topic it may also oscillate as it is in resonance with the more massive inner planet as it orbits the star, so would not be truly tidally locked to the star.

    @Ronald January 10, 2014 at 6:33

    ‘Leaves the crucial question: when a planet starts its existence as a Neptunian and then becomes a ‘secondary’ terrestrial through erosion’

    Erosion of an atmosphere takes some serious doing, look at Venus it is still there, water however is depleted through photo-dissociation and H/He is removed quickly. I personally can’t see a primary H/He atmosphere forming in the first place due to the collisional heat of formation of the planet.

  • Ashley Baldwin January 11, 2014, 13:47

    The bottom line is that exoplanetary science is still in its infancy despite great making strides ,even in the last few weeks. There is still a probable observational gap for earth sized planets in the HBZ around suitably old F, G and K stars . Absence of evidence rather than evidence of absence. Its going to take the next generation of planet hunting technology to bridge that gap , probably over the next ten years with TESS, JWST and WFIRST-AFTA ( that will end up at L2 in the end you watch ,hopefully with a star shade in tow at a later date ) and assorted mega ground scopes. I think the next gen Echelle EXPRESSO spectrograph due on VLT in 2016 will really get things going . I just hope the gap from pre existing Kepler data review to TESS launch is plugged by Kepler K2.

  • ljk January 14, 2014, 13:13


    Very Low-Density Planets around Kepler-51 Revealed with Transit Timing Variations and an Anomaly Similar to a Planet-Planet Eclipse Event

    Kento Masuda

    (Submitted on 13 Jan 2014)

    We present an analysis of the transit timing variations (TTVs) in the multi-transiting planetary system around Kepler-51 (KOI-620). This system consists of two confirmed transiting planets, Kepler-51b (Pb=45.2days) and Kepler-51c (Pc=85.3days), and one transiting planet candidate KOI-620.02 (P02=130.2days), which lie close to a 1:2:3 resonance chain.

    Our analysis shows that their TTVs are consistently explained by the three-planet model, and constrains their masses as Mb=2.1+1.5−0.8M⊕ (Kepler-51b), Mc=4.0±0.4M⊕ (Kepler-51c), and M02=7.6±1.1M⊕ (KOI-620.02), thus confirming KOI-620.02 as a planet in this system. The masses inferred from the TTVs are rather small compared with the planetary radii based on the stellar density and planet-to-star radius ratios determined from the transit light curves.

    Combining these estimates, we find that all the three planets in this system have among the lowest densities yet determined, ρp≲0.05gcm−3. With this feature, the Kepler-51 system serves as another example of low-density compact multi-transiting planetary systems. We also identify a curious feature in the archived Kepler light curve during the double transit of Kepler-51b and KOI-620.02, which could be explained by their overlapping on the stellar disk (a planet-planet eclipse).

    If this is really the case, the sky-plane inclination of KOI-620.02’s orbit relative to that of Kepler-51b is given by ΔΩ=−25.3+6.2−6.8deg, implying significant misalignment of their orbital planes. This interpretation, however, seems unlikely because such a event that is consistent with all of the observations is found to be exceedingly rare.

    Comments: 9 pages, 4 figures, accepted for publication in ApJ

    Subjects: Earth and Planetary Astrophysics (astro-ph.EP)

    Cite as: arXiv:1401.2885 [astro-ph.EP]
    (or arXiv:1401.2885v1 [astro-ph.EP] for this version)

    Submission history

    From: Kento Masuda [view email]

    [v1] Mon, 13 Jan 2014 15:56:52 GMT (233kb)