≡ Menu

Direct Spectrum of an Exoplanet

Astronomers have obtained a direct spectrum of the exoplanet HR 8799 c, about 130 light years from Earth, and if you watch your definitions, it’s possible to call this the first ‘direct spectrum’ of such a world. I throw in the qualifier because way back in 2004, astronomers using the ESO’s Very Large Telescope and the infrared instrument NACO obtained an image and a spectrum of a planet of about five Jupiter masses around a brown dwarf. The question then involved how the two objects formed — did they form together, like a stellar binary, or did the smaller object form out of the disk around the brown dwarf?

Whatever the case, the new work on HR 8799, also conducted with the VLT and NACO, takes us into interesting territory. Up until now, the way we’ve obtained a spectrum from an exoplanet has been to observe the planet moving directly behind its host star. The spectrum was then derived by comparing the light from the star before and after this event. That method relies, of course, on the orbital plane of the planet being aligned along our line of sight so that the ‘exoplanetary eclipse’ is observable from Earth, a significant limitation on our observations.

Direct observations like this one don’t rely on the planet’s orbital orientation, but they do count on remarkable adaptive optics for such ground-based work as astronomers tease out a planetary signal that is thousands of times dimmer than the star. The spectral information thus derived should offer clues to the planet’s formation and composition, according to Markus Janson, the lead author on the paper reporting the findings. Says Janson:

“The spectrum of a planet is like a fingerprint. It provides key information about the chemical elements in the planet’s atmosphere. With this information, we can better understand how the planet formed and, in the future, we might even be able to find tell-tale signs of the presence of life.”

The planetary system around HR 8799 is an intriguing one. The star is an A5-class about 1.5 times as massive as the Sun. Three planets are known here. HR 8799 b, c and d have estimated masses of 7, 10 and 10 Jupiter masses respectively, and separations of 68 AU, 38 AU and 24 AU. Moreover, we have evidence of a debris belt analogous to the Edgeworth/Kuiper belt outside of HR 8799 b, and a second debris belt similar to our own asteroid belt outside of HR 8799 d. The paper on this work refers to a ‘scaled-up version of our own Solar System,’ but it’s also a young one at an estimated 60 million years.

“Our target was the middle planet of the three, which is roughly ten times more massive than Jupiter and has a temperature of about 800 degrees Celsius,” says team member Carolina Bergfors. “After more than five hours of exposure time, we were able to tease out the planet’s spectrum from the host star’s much brighter light.”

Image: By studying a triple planetary system that resembles a scaled-up version of our own Sun’s family of planets, astronomers have been able to obtain the first direct spectrum of a planet around a star, thus bringing new insights into its formation and composition. The spectrum is that of a giant exoplanet, orbiting around the bright and very young star HR 8799, about 130 light-years away. This montage shows the image and the spectrum of the star and the planet as seen with the NACO adaptive optics instrument on ESO’s Very Large Telescope. As the host star is several thousand times brighter than the planet, this is a remarkable achievement at the border of what is technically possible. According to the scientists it is like trying to see what a candle is made of, by observing it from a distance of two kilometres when it’s next to a blindingly bright 300 Watt lamp. Despite the power of the VLT’s extraordinary adaptive optics system, the spectrum of the planet appears very faint, but still contains enough information for the astronomers to characterise the object. In the spectrum, several artefacts from the instrument are seen, such as internal reflections, or “ghosts”, and diffraction rings. Credit: ESO/M. Janson.

Yes, and it’s a spectrum that’s not in agreement with current theoretical models, a result that may be explained by a more detailed explanation of dust clouds in the atmosphere, or else a chemical composition different from what had been assumed. From the paper:

The results therefore imply that a more detailed treatment of dust in the models is necessary – or perhaps, that non-equilibrium chemistry is involved… Non-equilibrium models are worth to explore as they predict large differences in the spectrum as function of metallicity. Thus, a better understanding of the spectral behavior in this wavelength range might lead to a determination of whether or not the planet is metal-enhanced or not, and thereby provide further clues to its formation.

And this is interesting:

In the future, further observations with NACO can yield a spectrum also of HR 8799 b and maybe d, and yield a broader coverage of HR 8799 c by optimizing the observation procedure. This will provide the opportunity for comparative exoplanetology within a single system.

Advances in spectroscopy have obvious implications for astrobiology as we try to understand the atmospheres of distant planets and look, eventually, for chemical biomarkers. Thus the better we become at spatially resolving a planetary spectral signal from that of its star, the closer we are to reaching our long-term goal of identifying life elsewhere in the universe. The paper is Janson et al., “Spatially resolved spectroscopy of the exoplanet HR 8799 c,” in press at Astrophysical Journal Letters (available online).


Comments on this entry are closed.

  • Daniel January 14, 2010, 10:21

    i was check the extrasolar planet encyclopaedia (EPE) ,and there is 5 extrasolar planet that NEVER been announced…


    and this is scientific paper of 08/2009: http://exoplanets.org/papers/sixpack.pdf

    and this is extrasolar planet catalog in alphabetic order: http://exoplanet.eu/catalog-all.php?&munit=&runit=&punit=&mode=1&more=

    i send a e-mail for EPE, but they just told that the authors of this paper,don’t give the EPE the right of publicate this paper on the EPE

    this is strange because this planets it’s not secret, it’s already been announced since august of 2009 on the California Planet Search http://exoplanets.org/papersframe.html

    i don’t know, but for me, look like if something like that start happen in the EPE, that is the great reference in extrasolar planet of all.Now they could lose the credit of announce new exoplanets

    now because of this it’s supposed to have 429 planet and not 424,anyone know why,this planets it’s not on EPE catalog?

    the unannounced planet are: HD34445b, HD126614b, HD24496b, HD13931b, Gl179b

    there is one small note on the discussion on the end of this paper that talk about of the unannounced planetary system of Gl 179 a M dwarf star ,on a recently announced planet around HIP 79431 another M dwarf star with a gas giant planet see at:


  • Darrell E January 14, 2010, 11:49

    The first sentence of your post gives the distance to this system as “ten” light years, but in the image description the distance is stated to be “about 130 light-years away”. If you have a chance could you please clarify this?

    Very exciting work! I really wish that we had the ability to unambiguously determine what the inner reaches of this system look like.

  • Administrator January 14, 2010, 14:27

    Darrell, right you are. I hate it when I do that — somehow plugged in a completely wrong number there, and not sure what happened. It’s fixed now, though, and thanks for catching it.

  • Didac January 14, 2010, 15:50

    Daniel, more sooner than later, counting all discovered exoplanets will be a task as difficult as counting all small bodies in the Solar System. So, if the problem is only 5 out 429, I think we can cope with that. Moreover, the level of confirmation of all those planets is not uniform.

    As for the 10 light-year error, it is very obvious that if HR 8799 was at that distance, that star had a Greek or Arabic name.

  • kurt9 January 14, 2010, 18:16

    Its going to be a lot harder to do spectral analysis of an Earth sized planet than that of a 10 times Jupiter sized planet. I still think a space-based platform will be necessary for this.

    Nice solar system, thought. Two Kuiper belts many times larger than our own. Definitely rich in resources (for an O’neill type civilization).

  • Daniel January 14, 2010, 19:07

    Didac said:

    “Daniel, more sooner than later, counting all discovered exoplanets will be a task as difficult as counting all small bodies in the Solar System. So, if the problem is only 5 out 429, I think we can cope with that. Moreover, the level of confirmation of all those planets is not uniform”

    if you watch english football championship for example and you want know how many points for example livepool did, and TV and other communication way, don’t inform you how many points livepool or any other team did, how you get know was the champion on the end of the championship??

    information it’s to be divulgate

    this only “5” planet make a great diference on statistics extrasolar planet evolution speciality for example the forgettable planet around the M dwarf Gl 179 which is very important to study the formation of gas giant around M Dwarf stars and why this star have so few gas giants…

    and in this paper there is 1 planet that is in a binary system, and other it’s in a triple star system and one of this stars in this triple star system it’s only 33 ua from the primary star that host a planet a 2,35 ua unit… that is very interest for study about planet formation around a multiple star system

    and there is second thing EPE it’s great refence exoplanets,that way EPE start lose credit…

    look because this this wasn’t announced by EPE, try to find any reference of this 5 planets on Google,you go to find nearly none, i just find one refence of this planets in Russian, i don’t find refence of this planet ever in Simbad! and that is all because EPE not announced.

    this was i failed from EPE part,that is main source of information for the public and scientist

  • andy January 14, 2010, 20:28

    Daniel: so let me get this straight: you ask Jean Schneider about why these planets aren’t listed on EPE, and Jean Schneider tells you that it is because the authors of the paper haven’t given permission yet. Having been informed of this lack of permission for publishing details of these planets, you then go and post them on the internet in blog comments, and blame Jean Schneider for not being reliable.

    If there’s any problem here, it is surely not Jean Schneider’s fault…

  • Paul Titze January 14, 2010, 21:07

    It would help if we had a big fully remote controlled telescope in a crater on the Moon, any plans for this in the near future?

  • Daniel January 14, 2010, 23:37

    andy i’m can’t say if that is Jean Schneider fault, but the information of this 5 planets it’s free already by the California Planet Search, and if this information it’s free i can post, if EPE can’t it’s not my fault…

  • Ronald January 15, 2010, 8:12

    ref. Paul Titze: “It would help if we had a big fully remote controlled telescope in a crater on the Moon, any plans for this in the near future?”

    I have been thinking about that myself as well. One often reads about the need for an interferometer and/of coronagraph (TPF, Darwin) in space, in order to be able to derectly image and spectroanalyze earthlike planets.

    However, I would think that a large telescope (series) on the moon might be able to do at least the same, because of the near-total lack of an atmosphere.

    Even more tantalizing question is now, whether and to what extent the same might even be achieved on Dome A and/or C on the Antarctic Plateau.

  • Darrell E January 15, 2010, 10:26


    There is no atmosphere in interplanetary space either, so a moon based scope will not derive any advantage over space based scopes on that front.

    The distance between the occulter and the scope for a stellar coronagraph is pretty large, 72,000 kilometers for the Starshade design. Trying to use such an occulter with a moon / earth / planetoid based scope would present a very difficult and inefficient alignment exercise.

    I think that logistical concerns will be a very large factor in determining whether the best place for a given scope is low earth orbit, lunar surface, or deep space. But the equation will change over time. If we ever develop the moon to the point that people are living and working there full time, with a local industrial infrastructure to support them, then building and maintaining scopes on the moon would be quite easy. But, under those same circumstances it would also be much easier to build, launch, and maintain deep space instruments from the moon as well. So in that happy future logistics will be a much smaller part of the equation.

  • Eniac January 15, 2010, 13:02

    Why would we install a telescope on the moon, where it will be weighed down, threatened by dust and heat and cold, and hard to point at a fixed spot? Why go through the extra expense and risk of a moon landing? What is wrong with a telescope floating in space?

  • NS January 15, 2010, 14:44

    If I understand this right, HR 8799 c orbits at 38 au and yet has a temperature of 800 C? Is that because of the extreme brightness of the host star? I’m fuzzy on the various stellar types and what kind of planetary temperatures they’d give rise to.

  • andy January 16, 2010, 5:40

    NS: When a gas giant forms it is going to be quite hot. HR 8799 is a relatively young system, the gas giants haven’t had much time to cool down yet. The main contribution to Jupiter’s temperature is internal rather than the Sun, and Jupiter is much older and less massive than the HR 8799 planets.