≡ Menu

Water Vapor Detection on a ‘Super-Earth’

We’re beginning to probe the atmospheres of planets other than gas giants, a step forward that the next generation of space- and ground-based instruments will only accelerate. This morning we have word that the habitable zone ‘super-Earth’ eight times as massive as Earth orbiting the star K2-18 has been found to have water vapor in its atmosphere, making it the only exoplanet known to have water as well as temperatures that could sustain that water as a liquid on the surface. This is also our first atmospheric detection of any kind for a planet orbiting in the habitable zone of its star.

Angelos Tsiaras (University College London Centre for Space Exochemistry Data) is lead author on this work, which appears today in Nature Astronomy:

“Finding water in a potentially habitable world other than Earth is incredibly exciting. K2-18b is not ‘Earth 2.0’ as it is significantly heavier and has a different atmospheric composition. However, it brings us closer to answering the fundamental question: Is the Earth unique?”

Image: UCL’s Angelos Tsiaras. Credit: University College London.

If Dr. Tsiaras’ name seems familiar, it’s because you’re recalling his work on the super-Earth 55 Cancri e, reported in these pages back in 2016 [see Light, Dry Atmosphere of a ‘Super-Earth’]. Tsiaras and team used a methodology they developed to detect significant amounts of hydrogen and helium in this world, working with transmission spectroscopy data from the Hubble telescope. Now, collaborating with UCL colleagues, Tsiaras again uses Hubble, to analyze starlight filtering through the atmosphere of K2-18b as it crosses the face of the star as seen from Earth.

The molecular signature of water vapor is clear, as is the evidence for both hydrogen and helium, while such molecules as nitrogen and methane, if present, are undetectable at these levels. It will be fascinating to see whether we can move on from these observations to estimate the planet’s cloud cover and percentage of atmospheric water. The star is a cool red dwarf 110 light years away in the constellation Leo. And it presages the kind of work that the study of such nearby stars will generate, as co-author Ingo Waldmann (UCL) notes:

“With so many new super-Earths expected to be found over the next couple of decades, it is likely that this is the first discovery of many potentially habitable planets. This is not only because super-Earths like K2-18b are the most common planets in our Galaxy, but also because red dwarfs — stars smaller than our Sun — are the most common stars.”

Thus we move on to the next generation, which includes the James Webb Space Telescope as well as the interesting ARIEL (Atmospheric Remote-sensing Infrared Exoplanet Large-survey), slated for a 2028 launch by the European Space Agency. ARIEL’s charter will be to observe at least 1,000 known transiting exoplanets, going to work on their atmospheric composition, chemistry and thermal properties. K2-18b will obviously be one of its targets as we dig further into conditions on its surface. Meanwhile, we can expect TESS (the Transiting Exoplanet Survey Satellite) to detect hundreds more super-Earths as it continues its mission. No shortage of targets!

Image: Exoplanet K2-18b. This artist’s impression shows the planet K2-18b, its host star and an accompanying planet in this system. K2-18b is now the only super-Earth exoplanet known to host both water and temperatures that could support life. UCL researchers used archive data from 2016 and 2017 captured by the NASA/ESA Hubble Space Telescope and developed open-source algorithms to analyse the starlight filtered through K2-18b’s atmosphere. The results revealed the molecular signature of water vapour, also indicating the presence of hydrogen and helium in the planet’s atmosphere. Credit: ESA/Hubble, M. Kornmesser.

K2-18b was detected in 2015 through light curve analysis made possible by Kepler in its reconfigured K2 phase, its mass later constrained by radial velocity data from HARPS, leaving researchers to believe it is either a large, rocky planet or a water planet with an ice crust. The latter work, led by Ryan Cloutier (University of Toronto), also discovered a second, non-transiting super-Earth in the system, moving on an orbit interior to K2-18b.

The paper is Tsiaras et al., “Water Vapour in the Atmosphere of the Habitable-Zone Eight Earth-Mass Planet K2-18 b,” Nature Astronomy 11 September 2019 (abstract). Also see Benneke et al, ”Water vapor on the habitable-zone exoplanet K2-18b,” submitted to Astronomical Journal (preprint). Björn Benneke (Université de Montréal) is a co-author on the Tsiaras paper. [My mistake, looking at the wrong abstract].

The Cloutier paper referenced above is “Characterization of the K2-18 multi-planetary system with HARPS: A habitable zone super-Earth and discovery of a second, warm super-Earth on a non-coplanar orbit,” Astronomy & Astrophysics Vol. 608, A35 (5 December 2017). Abstract.


Comments on this entry are closed.

  • Andrew Palfreyman September 11, 2019, 14:26

    You might be interested to learn that the star is 111 lightyears distant.

    • Paul Gilster September 11, 2019, 15:01

      I’ve likewise seen it cited at 110 light years, which is why I used that figure; also, given the degree of uncertainty in these matters, rounding at 110 seems reasonable.

    • Andrew LePage September 11, 2019, 15:19

      I’m not sure about where the distance values are coming from but according to the DR2 Gaia parallax measurements, K2-18 is 124.2 +/-0.3 light years away. Here is a link to this object on SIMBAD:


      • Paul Gilster September 11, 2019, 17:23

        I’ll lean to Gaia, then.

        • Harry R Ray September 12, 2019, 9:32

          This also means that the planet’s radius is 2.71 Earth radii(much more akin to Gliese 1214b instead of Kepler 22b)instead of 2.37 Earth radii as posted on Abel Mendez’ PHL Upra website. Thus, it is clearly a CLASSICAL sub-neptune instead of a super-earth. This is precisely the TYPE of planet NEAR is most likely to have imaged around either Alpha Centauri A or Alpha Centauri B should a currently imageable planet exist inthe Alpha Centauri system. We will know whether one does or not by the end of next month.

  • Andrew LePage September 11, 2019, 15:12

    I did one of my earliest “Habitable Planet Reality Checks” on K2-18b (then known as EPIC 201912552b)


  • Daniel Högberg September 11, 2019, 15:20

    Interesting indeed! But planetary atmospheres aside, what will the gravity and atmospheric pressure be like at the surface of a planet with 8 times earths mass? Is mass, size and gravity somehow connected or does it not work like that?

    • FrankH September 13, 2019, 12:08

      A quick and fairly accurate way to estimate a planet’s surface gravity (compared to Earth’s) is to multiply the ratio of the radii (or diameters) by the ratio of the density (compared to Earth = 5.514gm/cm^3). So if we assume that this *sub-Neptune* planet has the same density as Neptune (1.638gm/cm^3) and 2.71 earth radii, we get:

      2.71 radii x (1.638gm/cm^3/ 5,514gm/cm^3) = 2.71 x 0.297 = 0.80 Earth’s surface gravity.

      If we assume a density equal to Earth, the answer is simple: 2.71g
      The answer will probably be somewhere in between those extremes.

      We don’t know anything about the conditions on sub-Neptunes, but if it’s anything like our ice giants, it’s going to have a crushing atmospheric pressure on its “surface” – probably a very hot ocean.

    • Robert September 13, 2019, 13:09

      Yes they are related. Surface gravity is MG/R^2 and for earth that’s 9.8 meters/second^2. G is the universal gravitational constant, M the planets mass and R the radius. So for this new world if we put in M and R as multiples of the earth values we get the ratio of 8.6/2.71^2 times g, the earth gravity at the surface which works out to 1.17 g or about 11.5 meters/second^2. Assuming this world has a solid surface you can walk on, you would only weight 17% more. You’d get used to it.

  • Jean Schneider September 11, 2019, 15:32

    I do not see B. Benneke as co-author of the Tsiaras et al. paper.

    You could make clear that Tsiaras et al. have used the Benneke data
    and that Benneke has published his analysis before Tsiaras (Accepted
    in Astron. Journal).
    It is the second time that one of the co-authors of the Nature
    paper publishes under its name the detection of water from
    the HST data belonging to somepne else (for the planet HD 29458 b).

    • Paul Gilster September 11, 2019, 17:23

      Yes, my mistake Jean, and thank you for catching it. I had the wrong abstract in front of me when I saw the Benneke reference. I wondered what was going on as the story seemed to be getting reported in entirely different ways. Thank you for clarifying this.

  • Alex Tolley September 11, 2019, 15:39

    If I understand the paper correctly, we have a world that has a thick H/He atmosphere with H2O vapor. The JWST, if successfully launched, can also detect CH4, CO2, CO, and NH3, but not O2. If the JWST detects CH4, CO4, but not CO, this would imply a fairly unambiguous detection of an atmosphere produced by methanogens, and therefore the presence of life. Pity we cannot detect O2 as that might allow for aerobic life supported by some version of photosynthesis.

    Tempering this, we know that Neptune has CH4, NH3, and H2O in its atmosphere, so it is possible that this world is like a warm Neptune. Could life appear on such a world? It would be fascinating if it did.

  • Geoffrey Hillend September 11, 2019, 16:56

    An exoplanet with eight Earth masses has the gravity to trap a thick atmosphere. A spectral signature of hydrogen and helium might indicate K2-18B is more like a mini Neptune. A spectral signature of methane would not be a surprise. Oxygen would be a surprise.

  • Mike Serfas September 11, 2019, 18:15

    When I was a child, I thought of astronomy as the dullest of sciences: there were nine planets, eight of them were dead, and nobody might ever know if there were more. Now you have some ground-breaking advance to report every day! Even if it is only a demonstration of water vapor detection, this is a key step forward.

    Apart from the technical aspect, it surprises me how many people forget the water rain in our own system. A planet like Saturn should have a lovely cloud layer at Earthlike temperature and pressure and a gravity only a few percent higher than that of Earth. It has too much hydrogen for any biochemistry to give it an oxidizing atmosphere, and there is a lack of a rocky surface and maybe of most trace minerals (I don’t know). I would love to see more about whether there are derivatives of polythiazyl or other polymers of S/N/C/O/H/P floating around in that layer. (I never did spot a followup report about the final data Cassini sent from the lower atmosphere – I suppose someone here knows!) Nowadays I am a little bit open minded about whether most of the life in this system is on Earth.

  • Steve Muise September 11, 2019, 19:28

    I’m pleasantly surprised by this news; I would have expected the lighter elements to have been stripped from the atmosphere by flare activity. Perhaps there’s a mass limit above which this becomes more improbable?

  • Martin September 11, 2019, 20:35

    Just to put things into perspective (at least in my mind):
    Given a mass of 8.6 Mearth and a radius of 2.71 Rearth I get a surface gravity of 1.2 (Okay, 1.17) g’s and a density ot 2.4 (2.38) g/cm3 The corresponding values for Neptune are 1.14 (I presume at the cloud tops) and 1.638 g/cm3. This planet is significantly less massive and more dense than Neptune and “stellar wind erosion” of its atmosphere doesn’t appear to present a problem here (its sun is an M3 red dwarf).

    • wdk September 15, 2019, 22:18

      Hello, David.
      Your numbers seem like a good point of departure. At the very least, they suggest a planet where we could stand on a surface perhaps without a cane. I believe the stellar radiation was a little higher than terrestrial.
      If the sun’s surface temperature were about 5800 Kelvin, expanded to 1 AU it would be about 400 Kelvin. Just to make life simple it can be used. If 1 earth radius and 1 earth mass gives an escape velocity of 25,936 fps ( quaint?), then we can expect something square root of 8.3/2.71 times faster: 1.75 x or 45,390 fps (13.835 km/sec).
      Now at about room temperature the speed of sound in our atmosphere in English units is about 1116 feet per second based with mass about
      29 units vs. atomic hydrogen (1) and helium (4). If we say room temperature is about 300 Kelvin, we can presume that an active ionosphere will have temperatures in excess of this value if the M3 star gets angry. How angry? What does a 1000 degrees Kelvin give us?
      Speed of sound is proportional to sqrt (gamma R/mass Temperature).
      Monatomic materials would have a gamma of 1.66; diatomic, 1.4.
      But the latter case, I don’t think it would be meaningful. But for 1000 degrees K and species H we get sqrt ( 1.66/1.44* 1000/300 * 29/1)
      times particle velocities vs. speed of sound for H in this environment.
      In English units that sound speed is about 1116 fps …? 11,886 fps
      for the first 1000 degrees Kelvin. Or about 3,623 km/sec.

      This is all approximate. because molecules have a bell curve distribution, so to speak. Meaning that some portion of the curve would be escaping of a given population. Someone mentioned that 1/5, perhaps that speed of sound at temperature vs. escape would be an indicator of a likely species depletion over eons.
      Atomic hydrogen would go if the exosphere were continually over 1ooo degrees Kelvin. It’s not like that here, but maybe there – or worse.
      But continuing in the vein you started, we can elaborate the picture…

      • wdk September 16, 2019, 11:22

        The entry above needs at least one correction.
        A comma was entered where I needed a decimal point.
        3.623 km/sec.

  • David September 11, 2019, 21:55

    Right very exciting and we have a chance to launch the first generation interstellar flybys.
    And we have this Dark Energy in Cosmology.

  • wdk September 12, 2019, 11:53

    Not only is this an exciting discovery, but the methodology suggests that the survey samples will continue with an as yet undetermined number of similar results. In other words, water vapor can be detected in atmospheres of local exo-planets. And here is the first case.

    But given that we see something emerging out of the shadows, here is
    some rudimentary examination.

    With habitable zone position and intermediate mass and density between Earth and Neptune, I would suspect that internal structure is still open to conjecture, depending on how large of a core this world contains and what the principal constituents are. Also, our solar system giant planets revolve in 12 hours or less, sporting considerable angular momentum. You would have to wonder whether that would be dissipated with a 36 day orbit around an M3 dwarf – and wonder how long that has been going on. Circulation of the fluid elements would
    certainly be affected by that. In addition you have convection regimes to consider. Also, would wonder if this world has any satellites. I suspect that outer planet satellites affect circulation in their primaries greatly, Neptune and Jupiter being examples.

    If you have an atmosphere as turbulent as Jupiter’s, for example, it is hard to image anything like a floating reef on which structures could form. But if flows were less turbulent at large scales, maybe the habitable zones could support some sort of floating reefs…
    Given that, then the paths of living organisms could go off in any manner. It would be a different type of ocean with different types of shores, perhaps distinguished by depths in bars. Interior heat could come up radiantly, but convective outflow could make a wreck of things.

    With a density of around 2 gm/cc3 and out beyond the snow line – which this planet isn’t, we already have examples of moons which are mixes of ice and rock. Some are thought to be more evenly layered than others, depending on their histories. And some have oceanic layers below their solid ice ( or ice and rock) surfaces. We would not expect
    layers of the same chemical nature here, but would there not be some
    uncertainties about how deep the depths? It would be nice to think that generations hence, someone could land there and open a hatch needing
    only to have to deal with a surface pressure of about 10 bar. If this object turns out to have a lot of H2 and Helium and the rest being traces, dirgible flight or living quarters might need to rely on heated air. Every part per million of CO2, N2, etc. would help.

    • Mike Serfas September 12, 2019, 18:15

      Has anyone modelled the strategies for life in turbulent atmospheres? To my ears, turbulence sounds like an energy source for flight. Drop a feather into a patch of air, and if it has a way to decide which way it thinks the air it is in is headed, it could try to sacrifice a little height to glide rapidly out of a downdraft and then latch its barbicels together to soar serenely on an updraft. But this is imagination, not a simulation.

  • Ashley Baldwin September 12, 2019, 13:30

    This planet has retained a substantial and primordial atmosphere – plus water – within the erstwhile habitable zone of a “moderately active” M dwarf. This is significant in its own right. At just 0.15 AU, it is still very close to its star and should provide reassuring proof that despire flares, X-rays and EUV, atmospheres are not invariably stripped away from at least some red dwarf habitable zone planets. And not a particularly massive one in this case – well within an order of magnitude of the Earth. If this planet had been stripped it would have shown up as a barren Super Earth rather than a soppy mini-Neptune. After recent publications to the contrary this is good news.

  • andy September 12, 2019, 15:55

    Well sadly it looks this is going to provide an interesting case study of the reporting of scientific results. The ongoing discussion around this one in terms of what claims are being made and who’s making them is definitely worth reading. (Plus as an added bonus it looks like there was some “drama” between the two teams involved).

    There’s an interesting discussion about the conditions on planets with hydrogen envelopes in this thread. Not exactly what I’d call “habitable” despite what (some of) the researchers are saying…

  • Geoffrey Hillend September 12, 2019, 18:04

    One has to take into consideration the density since Neptune has a surface gravity not much more than Earth. Jupiter’s atmosphere has some water vapor in it. With the H and He in it’s spectra, I am biased in favor of the idea that it is a mini Neptune or gas covered surface based on our own solar system.

  • wdk September 13, 2019, 11:49

    Regarding K2-18b water signature in comparison to solar system objects, some matters cross the mind:
    1. For Jupiter and beyond, we know that there are water clouds at some
    depth, but is it that we detect them from Earth – or that we infer their existence from atmospheric models, or something in between. After all, the outer layers of these worlds are below the snow line, but taking their adiabatic lapse rate, shall we say, in reverse, there is bound to be a temperature and pressure state where water vapor would be a component. You might be able to see the clouds with a Voyager, Cassini etc., but their signature from Earth must be weak.

    2. Now given that the water vapor exists, what sustains all that chemistry in a predominantly hydrogen and helium atmosphere?
    Where is the oxygen coming from – or what sustains the water
    vapor clouds over a global ocean made mostly of something else?

    In small satellites in these regions, we are left with icy masses (H2O)
    surrounding cores or mixed with other materials. The hydrogen most likely never hung around long after formation. But if you have a planet with an atmosphere full of hydrogen free oxygen is going to combust rather quickly…

    So it leads me to two possible alternatives to test for:

    1)a deep gaseous layer where internal heat breaks up chemicals and they well back up with an opportunity to form into steam or water vapor.
    2) A genuine ocean providing vapor pressures at corresponding atmospheric temperatures.

    We say that the stratosphere is above the troposphere here on Earth – and there is a relative lack of clouds.
    But our troposphere would probably look more like that of Mars if there
    were no oceans. And this planet’s atmosphere does not look like that
    of Venus.

    A water vapor signature across 110 light years hardly rules out the prospects for oceans. Saying that, I can still imagine something like what happens in a tea kettle at any Earthly altitude. When vapor pressure and atmospheric pressure match you get boiling and convection in the fluid. But locally you would still have to have concentrations of water ( in this case) based on pooling to be of any consequence.

    Very mysterious place.

  • Geoffrey Hillend September 13, 2019, 17:00

    Astronomers and astrophysicists can detect water vapor in Jupiter’s atmosphere from Earth based telescopes, but only with high resolution spectroscopy in the blackbody infrared band, or IR thermal emission spectroscopy. I am ruling out oceans since water under one quarter to one half an Earth mass of H and He gas covering a layer of pressurized water or ice does not count as oceans. A mini Neptune or gas giant could potentially have a magnetic field which might account for the atmosphere and lack of solar wind stripping. Also the volume and density of the planet matter since K2-18b is half the diameter of Neptune, so the atmosphere of H and He might be a considerable fraction of the total mass of the planet, but with most of the mass composed of water and ice like Uranus and Neptune.

    I do like the idea of a hybrid planet which is a rocky large super Earth with a solid surface, but another problem to be considered is K2-18b is in the habitable zone with a warm temperature which means that we have to consider that H, and He are light gases so they can easily reach the one fifth of escape velocity and be propelled out of the atmosphere so even without solar wind stripping most of these gases will be lost over a long period of time. Consequently, I have to assume that only a gas giant could have H and He in it’s spectra. The question is can anyone detect the spectra of the gases H and He in our atmosphere from the distance of 110 light years and I don’t think we can since there is to little of it and the atoms are too far spread apart in the upper atmosphere or exosphere.

    • wdk September 15, 2019, 15:40

      This discovery does lead to other questions. Helium is straight forward in its analysis due to its inertness with respect to chemistry. But hydrogen could be detected or released in many forms. In the lower atmosphere it would be diatomic and allow for little free O2 to mix with it. But if it is detected in isolation, it could be an ion released from dissociation of other atmospheric species: H2, H2O, CO2, CH4, NH3… Such breakdowns we would expect from UV or flare activity. But higher layers of the atmosphere tend to break down into lower and lower atomic or molecular weights as well with ascent into an ionosphere or exosphere – at least locally. Means to calibrate the ratio of hydrogen to other atmospheric elements would help, but it looks like it’s not there yet.

      Another aspect of this would be ionospheric or exospheric temperatures. I am acquainted with models of the Earth’s which might have a mean temperature over a month related to the solar cycle ( e.g. from 500 to 1000 deg Kelvin) and then for brief flares, maybe going several hundred degrees higher. There are associated mean velocities for atoms associated with each species and temperature, but the new considerations are local escape velocity and the seasonal temperatures and the flare temperaturs. If flares are frequent, then there might not be much difference from the mean. Free atomic H should go first, of course.,,

      But considering weeks ago, we wondered if there would be any atmospheres on dwarf planets in the habitable zone, it seems like an
      identified planet with water vapor should have a fair chance of having a moderately deep and mixed atmosphere rather than one like Neptune.

  • Geoffrey Hillend September 15, 2019, 18:47

    Dwarf planets can’t have an atmosphere in a habitable zone since their size and gravity and low escape velocity would result in no atmosphere if they were in the life belt where the temperatures are high enough for liquid water like our Moon and the lack of atmospheres in the Jovian moons. Io is an exception because it experiences stronger tidal forces due to it’s closest proximity to Jupiter which heat Io’s interior, but it’s atmosphere is extremely thin.

    If you mean Earthlike atmosphere by “deep and mixed atmosphere.” I like that idea, but it seems to me that if I take in mind the previously written principles tend to not support the idea of an Earthlike atmosphere with H and He in it and a solid surface. If you can see H and He in a planet from the distance of 110 lightyears, then most likely there is a lot of it like a gas giant if we take the mass, density, volume and diameter into consideration. H and He at life belt temperatures have the energy to escape a planet of Earth’s size or Super Earth over a long period of time which makes me believe the H and He in K2-18b was from it’s proto planetary disk before the birth of it’s solar system. I take into consideration the greenhouse effect in an Earth like atmosphere which keeps it warm. Also there is the solar wind stripping. One does not have to worry about these with a gas giant with a large amount of H and He in it’s atmosphere so the solar wind stripping loss and kinetic energy atomic escape will be negligible or only a small percentage of the total H and He. This is only a hypothesis that can be tested when we have examined more atmospheres of all sizes of exoplanets. I still remain flexible on that idea, but I stick to principles because they usually accurately predict observations. This planet is clearly not Earthlike or a super Earth. Mini Neptune’s are classified between 2 and 4 Earth radiuses. Super Earth Wiki. Maybe a mini Neptune does have a solid surface, but I doubt it. I eagerly await the spectra of other Earthlike planets and super Earths. With more data on exoplanet atmospheres should help support the principles. In theory in order for us to know for sure we would have to send a space probe to exoplanets, but that technology still has not been developed for cheep missions. I don’t expect to have that happen in my lifetime. It seems more practicable to build larger and more sensitive land and space telescopes to get this information. I think a star shade for the JSWT would be invaluable for exoplanet spectroscopy.

  • wdk September 16, 2019, 11:37

    “Dwarf planet” – My slip of the tongue, so to speak, or so to write.
    What I meant were planets about Earth sized which were dwarfed by
    Neptunes and Jupiters. And I would agree that dwarf planets like
    Ceres would not have atmospheres in these circumstances.

    Still, what has been discovered so far about K2-18b makes it a very interesting case.

    Among the communities interested in exo solar planets, there are those who look at the engineering, concentrating on probes and propulsion.
    Then there are many of us who can now preoccupy ourselves with the planets detected or characterized. Whatever you first choice, it is still nice to be able to gather information on where future spacecraft might head. And even if launch is years or generations away, sifting through candidates has its benefits.

    Yes, larger telescopes or more telescopes directed at the problem of detecting and characterizing exoplanets, I’m for it. For long while, telescopes were directed at stars and galaxies. The first, to determine their nature: structure, evolution, etc.; the second to dip into cosmology. But there is meaningful science to exoplanets as well, even if our reach to them looks as unlikely as Pharoahs commissioning jumbo jets.