No, it’s *not* the smallest exoplanet found!

By Phil Plait | April 10, 2008 3:00 pm

The web has been buzzing over what everyone is calling the smallest extrasolar planet found, weighing in at about 5 times the mass of the Earth.

Problem is, that’s not the smallest exoplanet found, not by a long shot. That record is still held by three planets massing 0.02, 3.9, and 4.3 times the Earth’s mass, orbiting the pulsar PSR B1257+12.

This newly found planet, orbiting the red dwarf GJ436, is the smallest yet found orbiting a Sun-like star.

I’m not picking nits here. I’m trying to be careful. The news tends to focus on planets found orbiting stars like the Sun (that is, fusing hydrogen in their cores as the Sun does) because those planets are the ones in environments most like ours. Certainly, those ones resonate with us, because one of the goals of this search is to find not only a planet with roughly the same mass as Earth, but one on which the conditions may be ripe for life to form and evolve.

Trust me, I’m all for that! I’m excited, terribly excited, about this search. I was a part of it, briefly, years ago, and it means a lot to me. But I’m also an astronomer, and I think it’s important to remember that not only were terrestrial-massed planets found first, they were found a full three years before any other exoplanets were found orbiting Sun-like stars.

Pulsars are dead stars, the collapsed cores of massive stars that exploded. The three planets in the system found in 1990 are probably nothing like Earth; they either existed before their star went supernovae, in which case they got fried but good, or they accreted after the explosion and formed under very weird conditions that make them exotic. Either way, they’re damn odd (to quote Kirk), and so they don’t work their way into our minds as well as more familiar planets.

But still, their discovery was a huge advance in astronomy! Who could have thought planets could either survive, or form after, something as devastating and titanic as the explosion of an entire star? It rocked astronomy when they were found, and their existence is still something of a mystery. How did they get there? What are they like? What does a planet look like when it’s baked by intense X-rays and particle radiation for thousands of years?

I don’t mean to detract from either the search for exoplanets around Sun-like stars (again, big supporter of it) or people writing up the news of such (obviously, also a big supporter). I just want to make sure the record is clear. It pays not to gloss over details sometimes, because in many cases that’s where the real treasure and real wonder lies. I don’t want folks to forget our prejudice toward Earth-like (or solar system-like) planets, when the exotic and bizarre might get overlooked.

CATEGORIZED UNDER: Astronomy, Piece of mind, Science

Comments (40)

  1. Kullat Nunu

    Why everyone is saying it has been found? The existence of the planet is not as certain as news articles suggest. For example, the venerable Extrasolar Planets Encyclopaedia lists it still as an unconfirmed (preprint of the discovery paper appeared at back in January).

  2. andy

    It’s doubtful that the planets would stay in orbit after a supernova. A system will become unbound if the mass loss exceeds half the total mass of the system, a condition which is easily fulfilled by a supernova explosion (a ~10 solar mass star leaves a neutron star remnant of ~1.4 solar masses). Sure, a supernova “kick” in the right direction might save one planet, but saving an entire system (and ending up with the planets on nice near-circular orbits) is pretty implausible.

  3. Kullat Nunu

    Oh, and thanks for reminding the PSR 1257+12 planets. People tend to always forget them because they’re extremely inhospitable and un-Earthlike worlds.

    PSR 1257+12 B and C were the first certain planets discovered at the time of their discovery¹ and with the Moon-sized A still by far the weirdest and most exotic IMHO. Although detecting planets around pulsars should be relatively easy, only one planet (a gas giant) has been found orbiting a pulsar.

    Given that they couldn’t have formed before the progenitor star went supernova, their history must be truly unusual. The fact that PSR 1257+12 is a millisecond pulsar suggests that it wasn’t originally alone. Somebody must have spun it up. What happened to its companion?

    ¹) A few planets, such as the ones orbiting Arrai (Gamma Cephei) and Pollux were detected earlier, but confirmed much later.

  4. andy

    Kullat Nanu: one possibility is that tidal destruction of the pulsar’s companion star created the disc out of which the planets condensed.

  5. andy

    (sorry I can’t copy+paste usernames competently, apparently)

  6. A fellow Phil


    I have a quick question about these extrasolar planets we keep finding. Recently (sort of), Pluto got downgraded from planet to dwarf planet status amid a lot of hubbub. My understanding is that it was because somebody finally put their foot down and defined a planet as having three specific properties, of which Pluto satisfies only two:

    – Must orbit a star.
    – Must have enough mass that its gravity forms it into a sphere.
    – Must have swept its orbit free (essentially) of debris.

    From what I have read, our current methods of discovering planets involve star occlusion (or is it lensing?) and star wobble. So here’s my question: are these planets actually planets? Obviously they are orbiting a star, otherwise we couldn’t use our current methods to discover them. I would imagine we can even determine their mass, especially using star wobble. But how can we have any idea whether or not they’ve cleared their orbits? I would imagine that our discovery methods make it very difficult for us to find anything that’s relatively small, like orbital detritus. Are we just using “planet” as a convenient word, or do we actually classify these objects as planets? Am I just being picky?

    The other Phil

  7. Wait, 0.02 earth masses? I wasn’t aware our methods of finding exoplanets could see something anywhere near that small. Is there something I’m missing?

  8. Davidlpf

    The other phil, when actually visit these systems to see if the planets clear their orbits or found way of detecting anything else in the same orbit the best thing to call them are planets. The only things we can say for certain there are masses of such and such at so far away from the star, planet just sound a lot better.

  9. wisnij

    Wow, I can’t believe I hadn’t heard about PSR B1257+12 — that’s remarkable! Maybe in a few thousand years we’ll be able to get out there and do some close-range pulsar research. :)

  10. Either way, they’re damn odd (to quote Kirk)

    When did Kirk say this? (I presume you’re referring to James T. Kirk.) I remember him saying something like “That’s damned peculiar”, probably in Wrath of Kahn, but not “damn odd”.

    (Get a life? Why should I?)

  11. Cusp

    Yawn, planet finding is so passe.

  12. Jolly Bloger,

    The pulsar exoplanets were found by radial velocity techniques. For something like a pulsar that is very nearly a perfectly periodic radio source, you can get very precise results. Thus the moon-sized planet of PSR B1257+12.

  13. bc

    Fellow Phil,
    We have a good idea the exoplanets have cleared their orbits due to our understanding of orbital dynamics. The Exoplanets we have discovered tend to be dynamically like Jupiter (able to kick other small things near them out), due to high relative mass compared to their star and/or orbiting close to their star. Even the 5 earth mass planet orbiting Gliese 581 could be considered more like Jupiter in terms of dynamics because it orbits a star that is about 1/3 the solar mass at a distance of 0.073 AU, which is rather close.

    There are also systems that have a dust ring, although I think this is understood to be small dust particles glowing in infrared, and not objects as large as asteroids. Epsilon Eridani is I think an example of this.

  14. andy

    Actually, the pulsar planets were discovered by timing, not by radial velocity. The difference is that the timing measurements are made by observing how early or late pulses are received: as the planet orbits, the pulsar moves towards and away from us: when it is further away the pulses are received later (as it takes light longer to reach us), when it is closer, the pulses are received earlier. This method is more sensitive to planets on more distant orbits, since for a more distant planet, the radius of the star’s orbit around the centre of mass is larger, so the time difference is greater. The radial velocity used to detect most planets is most sensitive to planets on close orbits, since these make the star move faster.

    In fact, since the pulsar system has an inclination of about 50 degrees to the line of sight, the detection is effectively that of a 0.015 Earth mass planet, which at 0.19 AU from a 1.4 solar mass pulsar means the pulse times vary with a semi-amplitude of just over 3 microseconds. That kind of timing is incredible and only possible due to the remarkable nature of pulsars.

  15. PhxBruzer

    Wait a minute….. We found a Planet 5 times the mass of Earth going around a red dawrf star?? Can someone say Krypton? Superman can’t be far behind. 😉

  16. chip

    noticed you say the star went “supernovae.” my latin’s a little rusty, but is that plural? in other words, is a supernova a series of events, not just a single one?

  17. JB of Brisbane

    The Anderson TV series “Space: 1999” predicted that we would be finding and looking at extrasolar planets by the year in question. Where they got it wrong was in predicting that we would have a means to send manned expeditions to these planets, in time frames of less than a year.

    Back in the distant past (or at least when I was in high school) I remember reading that Barnard’s Star showed signs of having exoplanets. This may have been the reason that Douglas Adams mentioned it in The Hitchhiker’s Guide. Can Phil or anyone tell me if this was ever confirmed or otherwise?

  18. Sergeant Zim

    From what I read from the article, this planet was found by observing the ‘wobble’ of other planets around this same star.

    I get that, I also get how the orbit was calculated (sort of, my own math education stopped at calc. 1).

    What I don’t get, though, is how the rotational period was determined. Can anyone shed some light on this?

    I also disagree with some of the conclusions in the article about the requirements on a planet for it to be a home for life, specifically: “To sustain life, a planet must have a mass similar to that of earth, liquid water on its surface, an atmosphere and a similar orbital distance from its star as that of the earth from the sun.”

    I may be ‘way off base here, but wouldn’t a planet farther from it’s primary be able to sustain life, especially if it had more CO2, creating a higher greenhouse effect? Other variables in the ‘life equation’ may be just that- variable (it seems to me).

  19. Do we have metallicity data on Gliese 436?

    My out-of-date ISDb entry on this star, , has no heavy-element abundance data, meaning there wasn’t any that I could find back when I populated the database in the late 1990s.

  20. Tyler Durden

    There was also a 3 Earth Mass planet discovered in 1996 that was not orbiting a pulsar:

    Amazingly, it’s supposed to be over 7.8 [i]billion[/i] light years away.

    Unfortunately since it was discovered via gravitational microlensing there’s no way it can ever be found again to confirm the discovery.

  21. Cusp

    > There was also a 3 Earth Mass planet discovered in 1996 that was not orbiting a pulsar:

    Any mass estimate from a single event in high optical depth microlensing is likely to be complete boll*x.

  22. Thanks Barton and andy… I actually learned about the pulsar method back in school but had completely forgotten! Time to bring out the ol’ textbooks again.

  23. Douglas

    The other Phil:

    You mentioned: “My understanding is that it was because somebody finally put their foot down and defined a planet as having three specific properties,” and then asked about these exoplanets.

    In fact, the IAU Resolutions which you describe only refer to the definition of planets within our own Solar System (see for example the PDF with Resolutions 5 and 6 at ). Therefore, strictly speaking, the resolutions do not say anything about “exoplanets” one way or the other.

  24. Not only that, but the discovery is kinda doubtful so far. Greg Laughlin over at systemic wrote about it a while back and reserved judgement.

    Besides, Swiss exoplanet guru Michel Mayor passed by here yesterday and we got a little exclusive interview thing (in Swedish! YHBW) with him. He says his group’s measurements don’t show the changes in transit timing that you’d expect if Ribas and his colleagues’ model is right. So it’s a bit too early to say that there really is a super-earth hiding there.

  25. Steven Charles Raine

    Excellent news & well spotted with the pulsar’s planets being the first known from back in 1991. Still great to hear of another 5 earth mass planet being found (joining OGGLE 2006-BLG 390 or “Hoth” & Gliese 581c or “earth’s Not so-Close-Sister”) although I’m not so sure its necesarily even rocky in nature rather than a mini-Neptune … Or another “Hot Ice” world like the innermost of Gliese 436’s worlds.

    Truth is, I think we really do not know what a 5 earth-mass world is like…

    [b] Another Phil [/b] seems to have a very valid point – orbital clearance is just such a shockingly *_bad_* criteria – I think the IAU blundered massively in throwing out Pluto on such poor grounds.

    When we can’t tell for sure … & if something the size & mass of Neptune or Jupiter or above _isn’t_ a planet ..??? I mean, come on!

    I’ll also add one of the Pulsar planets – the outermost one has a mass of just 1/5th Pluto’s but still counts as an exoplanet and that by the idiotic IAU definition no star other than our Sun has a planet as strictly defined by them *only* the Sun can have planets. Argh! That was just so gosh-darned, frak buzzardedly dumb a decision of the IAU’s! :-/

    That first proposed definition :

    – spherical enough to be round by self-gravity,
    – orbiting a star directly &
    – not large enough to shine by hydrogen fusion

    To me that’s by far the best definition of ‘planet’ & I sure hope the IAU revisits its planets definition and revises it to something like that to correct its Prague meeting error.

    Meantime, I’m really looking forward to seeing the discovery of exoplanets in the 3-2 Earthmass range -and smaller! 8)

  26. Andy C

    I was pretty surprised when Phil said “0.02, 3.9, and 4.3 times the Earth’s mass”, my initial reaction was that he must have meant 3.9x and 4.3x Earth, and 0.02x Jupiter. You learn something new everyday.

    Interestingly, these are the only 3 planets at the JPL site that don’t have masses associated with them, does anyone know if this is just an oversight, or is there are a reason?

  27. Linusrp

    That first proposed definition :

    – spherical enough to be round by self-gravity,
    – orbiting a star directly &
    – not large enough to shine by hydrogen fusion

    Going by the last definition you would also include brown dwarfs as a planet? If something like this definition were to be used, it should perhaps be: – Not massive enough to have had Deuterium burning at some time during its excistence. And even that would leave the solar system with a heck of a lot of planets (around 12 at the moment, likely more to come).

    but I am digressing here…

  28. andy

    As for the fourth object in the PSR B1257+12 system, I have not seen anything which claims it is or isn’t a planet. The papers referring to it generally seem to describe it as cometary or asteroidal.

    Unfortunately the IAU definition is somewhat vague and doesn’t define the term “cleared its neighbourhood” very precisely. You can make an estimation of how well a planet will have cleared its orbital neighbourhood (apart from the 1:1 resonance, e.g. Trojans, quasisatellites) from the planetary mass and orbital distance: see this paper (PostScript format), which defines the likelihood that the planet can clear its orbital neighbourhood in a Hubble time. Applying the ?>1 criterion to the solar system, Pluto and Ceres do not fit the definition, the 8 major planets do. (Interestingly, the moon would also fit the criterion if it were in solar orbit, which suggests this criterion could be used for double planets) Applying the criterion to the PSR B1257+12 system, all three planets fit the criterion, the fourth (cometary/asteroidal) object does not.

  29. Cusp

    The issue of “what is a planet” is more yawn inducing than planet-finding itself.

  30. andy

    The issue of “what is a planet” is more yawn inducing than planet-finding itself.

    If as you imply you find planet-finding boring, why bother commenting on this post?

  31. bc

    The only thing I could think of more boring that finding planets would be finding extra terrestrial life. How passe.

    Anyways, I wanted to comment on the middlingly boring subject of determining what is a planet. I had a discussion with a astrophysicist tonight (I am traveling overseas, so it is night for me) and he mentioned that it looks as though there is a linear relation between star mass and planetary mass. That I think lends to the idea of a terrestrial planet of 5 earth masses around a M star being dynamically similar to Jupiter in the sense of “clearing the orbit”.

    I’ll try to report as accurately as possible what he said. I’ll start by saying that this is a guy who models planet formation and dynamics of exoplanets, so far more of an expert than I.

    Basically, his take is that dwarf planet is a meaningless term. This is because “dwarf planets” are in the same dynamical class with asteroids, KOB’s and comets, none of which are present in a mature system unless trapped there, such as the asteroid belt by Jupiter and Mars. As for the KBO’s and comets, that region of the solar system hasn’t reached full maturity, ie not enough time has passed for the m to be deflected or for a collision to occur.

  32. andy — you’re right, it was timing and not RV. My bad.

    JB — yes, there was a time when they thought they had detected first one and then two planets of Barnard’s star.

    There were two previous waves of exoplanet finds prior to the one we’re having now. They used astrometric techniques (wobbles in star sky paths, essentially). The first was in the 1940s, when Kaj Aa. Strand found an 8-Jupiter-mass planet of 61 Cygni, and I think a couple other such planets were found as well. The second wave was in the 1960s, when people working at Sproull Observatory in PA (Sarah Lee Lippincott, Peter Van de Kamp, and others) discovered a whole host of exoplanets, including one of Barnard’s star that was just 1.5 Jupiter masses in 1963. In 1968 the Barnard’s data was revised to a 1.6 Jupiter mass planet in a highly eccentric orbit, and the same year another interpretation came out which had two planets in coplanar, nearly circular orbits.

    People suspected these finds might be spurious. For one thing, the Sproull people kept finding planets with periods of about 24 years. In 1973, George Gatewood and a colleague I can’t remember offhand did an astrometric analysis of Barnard’s Star and found nothing. They also discovered that the lens plate and photographic emulsion had been changed at Sproull in the ’40s, just enough to be finding 24-year periods in the 1960s.

    All the early discoveries — 61 Cygni, Lalande 21185, Barnard’s Star, 18 Epsilon Eridani, etc. — are now considered mistakes. Not useless work, because astronomers were able to sharpen their techniques and learned how to avoid making certain types of mistakes. But all the findings were spurious. This is one reason there was so much suspicion of Mayor and his colleague’s 1995 find of the 51 Pegasi planet.

    You can still find a lot of old astronomy books which mention the planets of 61 Cygni and Barnard’s Star.

  33. Celtic_Evolution

    One of my favorite sites from way back,, used to have pretty useful information on exoplanets. I think the site was as much for entertainment value as anything else, but it did at least provide some rudimentary scientific information on known, suspected, and disproven exoplanets, all in a pretty well laid out format with very nice graphics and artistic renditions. But it looks like the site hasn’t been updated in a couple of years (since Sept. of 2005, in fact). More’s the pity. :(

  34. Celtic_Evolution

    I see Tyler Durden referenced the Extrasolar Visions site already earlier in this post… well… now you have two references to it. Get over it. :)

    I still like the way that site was laid out and presented and would love to see it taken over by someone who would actually keep it up to date.

  35. andy

    Celtic_Evolution: that was one of my favourite websites too. Unfortunately, the lack of updates is due to the death of the creator. :(

  36. Celtic_Evolution

    Nuts… what a terrible story. :(

    Well… I’d love to see someone take up the task of maintaining the site, if even as an homage to the man.

  37. Brett

    GJ 436c isn’t the smallest planet we’ve found around another Main Sequence star, either. Kepler found a Mars-sized planet around KOI-961, a red dwarf.

  38. Brett (38): Check the date on this post. :)

  39. Brett

    Doh! I can’t even remember how I found my way to this post, but I know it was some type of Google Reader Voodoo.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!


See More

Collapse bottom bar