Astronomers find most massive star ever discovered

By Phil Plait | June 7, 2007 4:42 am

This is really cool — astronomers using the Very Large Telescope (yes, I know the name sucks) in Chile and Hubble have discovered the most massive star ever found… and it’s a bruiser, weighing in at 114 times the mass of the Sun. This blows away the previously most massive known star, which had a mere 83 solar masses.

Cripes. That’s a whopping huge star. It’s labeled "A1" in the image above.

They were able to reliably determine its mass because it’s in a binary system, orbiting another star (actually, they orbit each other). The orbital period is — get this — 3.77 days! By carefully observing the Doppler shift of the stars’ spectra, they were able to find out not only the velocities of the stars’ motions in their mutual orbit, but also other orbital parameters (the ellipticity of the orbit, its period, and so on). These in turn can give you the mass.

Usually, the tilt of the orbit messes things up. If you observe it face-on, then you get different numbers than you do if it’s edge on. In other words, without knowing the tilt you can’t really know the mass. However, it was found that this system is an eclipsing binary: every orbit, the stars pass directly in front or behind each other from our view. That’s prefect! You can only get such a system if it is edge-on or nearly so. That constrains the tilt severely, and makes the masses easier to determine.

So in this case, we can be pretty confident of the mass of 114 times the Sun’s. I’ll add that the companion star is no slouch, either: it has 84 solar masses! I don’t know the error bars on these masses, so I can’t say for sure that even the companion star would have broken the old record. It very well might have.

114 solar masses is getting close to the limit on how massive a star can get. More massive stars fuse elements in their core far, far more quickly than lower mass stars. That generates mind-numbing amounts of energy, making the star incredibly luminous. At some point, the star gets so massive and so bright that to a particle sitting on the surface, the force of gravity down toward the star is offset by the force of radiation pressure (literally a pressure exerted by light itself) up, away from the star. The star literally is too bright, and launches its outer surface into space. This is called the Eddington limit, after the astronomer who first figured it out. The actual upper limit is hard to determine, even theoretically. It can depend on many things, such as how fast the star spins, and what elements are in it — oddly enough, the presence of the element manganese lowers the mass limit because it is very effective at absorbing visible light, which tends to make it more susceptible to radiation pressure. That’s the kind of incredibly detailed stuff that makes these calculations very tricky indeed.

In the image above, by the way, the star marked "C" is also a binary and thought to be another biggun, but the masses have not been reliably determined yet.

I keep thinking about this system: a pair of 114 and 84 solar mass stars, so close together they orbit one another in less than 4 days. They must be elongated stars; the centrifugal force would be ferocious, I’d wager, and the upper layers of those stars are barely holding on due to the Eddington limit. Unfortunately, no paper was linked in the press release I received, so I can’t go into too much detail without it being pure speculation. Are the stars blowing off a vast wind of material? Are they emitting gamma rays, like the scary massive binary in the cluster Westerlund 2? How long before these stars explode (and man, when they eventually explode, they’ll explode)? I wish I knew! I’ll have to keep my eyes open for the paper. Stuff like this is thrilling, just to know such an exotic and terrifying beast can exist.

I have to add this… the binary star is at the center of NGC 3603, one of my all time favorite nebulae. Check the banner at the top of this page to confirm this. :-) In fact, I have a whole page about the logo and the nebula. See those weird bluish rings around the bright star above the center of the press release image? The star is Sher 25, another crazy giant in that nebula that’ll blow up one day, and the rings look suspiciously like the rings I studied around a different star for my PhD. I considered observing Sher 25 with Hubble a few years back, until I realized it would take about 20 orbits of data — that’s a lot — and probably a year out of my life to reduce the observations and convert them into usable data. I decided not to go for it. Even now, I wonder if I should’ve tried. It would have been very cool.

Oh well. I love what I’m doing now. I get to write about stuff like this, and that sure is fun too!

Comments (61)

  1. Brian

    Thanks Phil. Is there something about the particular neighborhood the leads it to spawn so many giants? Are these low-metallicity stars?

  2. OneHotJupiter

    Wow! That’s a big ball of gas!!

  3. 84 was the previous largest star? I thought Eta Car was larger than that — somewhere between 100 and 150, if I recall.

    Also, is it possible that the two stars being so near each other might help each other “exceed” the Eddington limit?

  4. I’ll echo what Stephen said, and I thought it was believed that the “Supermassive Supernova” (SN2006gy)?

    But I’ll agree this is cool news.

  5. Eta Car’s mass isn’t well known. It is almost certainly near 100 solar masses, but it hasn’t been pinned down. The super-supernova progenitor was not seen, so we don’t know for sure how massive it was. It’s theoretical.

  6. Chip

    Woke up this morning, made coffee, checking my emails, then I read this. Orbiting each other in 3.77 days and at 114 and 84 solar masses.

    Excuse me, I think I’ll go hide under the covers.

  7. astro groupie

    Thanks for explaining the Eddington limit–now I know what all my astro friends are muttering about as they read this very piece of news!

  8. Is the Eddington limit still viable considering dark matter has entered the picture? I postulate a star with a dark matter core could be much more massive.

  9. Frank

    Any real possibility that these two stars could collide someday?

  10. MarshallDog

    From what I understand, I don’t think a star can have a dark matter core. Dark matter does make up the majority of the universe, but you can only really tell it’s there in larger scales. I know a star is big, but a galaxy is much bigger. Phil has said it many times… a cloud of dark matter could pass right through your body and you wouldn’t notice. I would guess the same would be true even for something the size of a star. But if a galaxy sized cloud of dark matter were to pass through the Milky Way, something might move.

    Then again, I don’t really know all that much about dark matter, so don’t listen to me.

    As for Stephen’s question about the Eddington Limit, I would think that two stars right next to each other would make it harder for each to break the limit because the matter trying to accumulate has twice the luminosity to overcome. I’m not sure if that makes sense to anyone but me.

  11. Gary Ansorge

    Doubled luminosity only applies on the facing sides of the stars. On the edges both would be experiencing the radiative effects of themselves and the other, so there may well be outgassing streamers from the edges.

    What’s their distance from us? Would their going nova be readily apparent?

    GAry 7

  12. Brian

    I am no expert, but let me test my understanding, subject to correction by those better informed. My understanding is that dark matter particles do not clump together because they lack a mechanism for radiating away energy (baryonic matter can radiate away energy in the form of photons). Thus, while a few of the less energetic dark matter particles might get trapped quite close to the center of mass of the system, the majority would not and would remain able to zoom about pretty freely. This freedom would prevent most of the dark matter from becoming bound into the core of a star.

  13. DrFlimmer

    This is incredebly cool! It’s absolutly fascinating! (I think I can find many more words like this! ;) )

    I just wonder how such hugh, no, such extremly massive stars can form. That cannot be done just by accreation, before all the mass has fallen into the star it would have blown it away! Maybe they are formed by collisions and these two stars are left and will collide one day (maybe bacause they are radiating away gravitational waves), but, well, maybe they will blow up before they can collide, maybe we can see that…

    An interesting field to study, no discussion!

  14. log

    Fascinating! I think my brains about to exlode.. 114 times mass of the sun!

  15. Sorry for the off-topic comment, but CSPAN-3 now has live coverage of a Congressional hearing on the NASA Inspector General, Robert W. Cobb. So far, nobody has had anything good to say about him.

  16. Ken G

    A couple interesting clarifications I can add:
    1) The Eddington limit only applies to the luminosity of a given star, not to that of its companion. The geometry is quite important to the limit– radiative flux from an external companion, by conservation of energy, both comes in and goes out, and so cannot alter the Eddington limit of a given star. Streamers off the surface do seem possible though, and certainly the winds of the stars would be very strongly altered– but not the basic stellar structure.
    2) Phil’s point:
    “More massive stars fuse elements in their core far, far more quickly than lower mass stars. That generates mind-numbing amounts of energy, making the star incredibly luminous.”
    … can be read incorrectly, fueling a common misconception you see in a lot of places. It is true that the luminosity of a main-sequence star is furnished by nuclear burning, but this is not what *sets* the luminosity– the logic is rather the reverse: the nuclear burning occurs at whatever rate it needs to do to furnish the appropriate luminosity. The appropriate luminosity, on the other hand, is determined primarily by the size of the star (how big is the “bucket” of energy) and the rate that the light escapes (how big are the “holes” in the bucket). Thus massive stars are highly luminous strictly because they are large– they are “big buckets” of energy, and the time it takes for the energy to escape is not all that different from smaller stars.

  17. Macz

    If there is such a huge number of stars, we should see stars colliding and exploding all the time. Have we ever observed 2 stars collide within a humanly observable timeframe? It seems statistically likely that we should be able to these types of events this if we look in enough places. We always seem to see them before or after, not during.

  18. MarshallDog


    There is a huge number of stars, but there’s an enormous amount of space between them, so the density of stars in the universe is actually very low. Consider the impending (13 billion years from now) collision of the Andromeda and Milky Way galaxies. While both contain trillions of stars, both are diffuse enough so that no stars will actually collide during the “collision”.

  19. My nose always drops off (I’m distantly related to Tycho Brahe (no, not really)) when I try to visualise the size of the sun compared to Earth, but 114 times the mass is just… unimaginable.
    How much larger than the sun is this then?

  20. OtherRob


    Is there some physical reason that “no stars will actually collide” when the Milky Way and Andromeda galaxies collide? Or simply the statistically unlikely possibility that such a collision will occur?

  21. John

    OtherRob, I believe it is just statistically improbable that it would ever happen. It would be like throwing two golf balls around a baseball stadium at random and having them hit each other. Poor analogy, but I think it might give you a visual idea :P

  22. DrFlimmer

    Forgot something:

    Mr Plait, if you find the “original” paper which has been or is about to be published, could you please link it? I would like to read it myself!

  23. john

    Isn’t VY Canis Majoris bigger than this?

  24. He won’t give the paper, but he can have the University of Montreal host his images for him. How nice that he doesn’t have to pay for bandwidth for images.

    How I wish this page had been Dugg instead of this. It gives legitimate blogs a bad name.

  25. Stark

    John, Good analogy… but the scale is way off!

    Take, for example, our Milky Way. A middle of the road estimate puts 100 billion stars in it – spread across a disc 100,000 lightyears in diamater and roughly 1000 lightyears thick. So we can do some back-o-the napkin number crunching and get some rough idea of stellar density in our galaxy.

    Volume of the galaxy : ~2.6 x10^51 Km3 (aka. REALLY BIG)
    Numbers o’ stars in the sky (ok, galaxy) ; ~100 billion.

    This gives us a star for around every 2.6X10^40 Km3. Ok, that’s a hard number to grasp so call it roughly a star for every 310 cubic lightyears or so. Staioll very hard to consider isn’t it? Considering an average star is all of 4.7 lighseconds in diameter…. well, you get the idea.

    So, scaling to golf ball sizes you’de get 2 golf balls in a space several times larger than the diameter of the solar system. It’s safe to say that the probability of stellar collisions in a galactic collison is very, very low. It’s true that this does not account for the fact that stars are more densely distributed in some areas than others… but still, it is a very low probablity event.

    While we see alot of stars in the sky the vast majority of the galaxy is empty space. Well, OK, it may be dark matter but for practical purposes it is empty space.

  26. Brian

    Thanks for the link.
    Many of us enjoy and benefit from Phil’s blog.

  27. Angie, I’m not sure what your point is.

    First, your claim is incorrect. The image displayed on my site is hosted on my site. It is a low res version that links to the higher resolution image on the Montreal site.

    Second, when I wrote this entry up, the press release was not on the Montreal site. I looked for it. There was therefore no way for it to be Dugg. Had they included the URL for the permanent location of the release in the embargoed press release I received days ago, I would have linked to it.

    Third, I do not control who Diggs what.

    Fourth, had the release been online when I wrote this, I could have simply linked to it and said “Look, a cool press release!” but then there is no insight, no color, and no personal relationship to the news. If people want the raw press release, they can go to any number of other sites.

    Fifth, I take your comment about “legitimate blogs” somewhat personally. I am a professional writer and author with a PhD in astronomy (in the very field dealing with this blog entry). Bad Astronomy, I am very proud to say, is the number one astronomy blog on the planet. So I’m curious about your insult. Would you care to elaborate?

  28. Steve P.

    Eh, if Angie had given BABlog anything more than a 5 second glimpse she’d realize it’s the baddest (not just astronomy) blog in the solar system. Not the galaxy, though, I reckon.

  29. I’m not an astronomer, but I’m confused: Are you taking about the Chandrasekhar limit here? He initially developed the idea, then Arthur Eddington discouraged Chandrasekhar in keeping this line of research (or at least this is what is writtent in Stephen Hawkins’ Brief History of Time. Hope I’m not making a fool of myself. Keep the good work, especially with debunking the neo-creationism movement a.k.a “Intelligent” Design (a link to the exposition in evolution at the Field Museum in Chicago wouldn’t hurt…).

  30. Ken G

    Carlos: the Chandra limit is for the mass of the non-nuclear-burning core of an evolved star before it collapses. It is not related to the luminosity/mass limit that is the Eddington limit.

  31. eh?

    I thought VY Canis Majoris (Humphreys model) or VV Cephei A was the largest star…

  32. darius


    Seriously, learn how to use the internet before making comments about bandwidth-stealing. It took me about half a second to determine that your statement was wrong and that BA was hosting the image on this page himself. Also, as BA said, your other comments were out of line.


    That Wikipedia entry is referring to diameter, BA is talking about mass. Though I’m far from an expert, I would say that there are large variations in density due to the actual compositions of various stars.

  33. Darth Robo

    “Orbiting each other in 3.77 days and at 114 and 84 solar masses.”

    Woof, them’s big! I think I’ll sit this dance out! (dizzy) Yikes!

  34. Buzz Parsec

    About dark matter being captured in giant stars…

    Consider 2 objects moving past each other. If their only interaction is gravitational, then they’ll pass right by each other (or even through each other!) without losing any energy. They’ll accelerate as they approach each other and then decelerate as they depart, but the relative speeds when they are very far apart will be the same before and after the interaction. To slow down and be captured by each other, they need to interact somehow and lose some energy when they are close to each other, and the whole distinguishing characteristic of dark matter is it doesn’t interact. (Or doesn’t interact very much, it might be weakly interacting, which means the particles have to be extremely close, like within the size of a neutron, for anything to happen!) So two dark matter particles or a dark matter particle and a huge star won’t combine. This is why dark matter has very little tendency to clump.

    However, with 3 massive objects, everything changes! It also gets incredibly complicated, way beyond my celestial mechanics skills. With 2 massive stars and a dark matter particle, maybe it could get captured. Or maybe it just increases the chances enough that it would take trillions of years instead of quadrillions of years to capture a dark matter particle. Anyone know?

    BTW, “the scary massive binary in the cluster Westerlund 2″ is a great name for a science fiction story. Or a rock band.

  35. Tim G

    What is the luminosity of this thing? Going by the rule of thumb that the luminosity of main sequence stars scale by a factor of 3.5 with mass yields a solar luminosity of over 15 million. I suspect that this rule of thumb is overextended and I may be off by a few orders of magnitude.

  36. OtherRob

    John and Stark,

    Thanks for the info. And the mental image of a couple of guys heaving golf balls around Turner Field. :-)

    The idea of two galaxies colliding without ever actually touching each other is just mind blowing. I love this stuff. :-)

  37. Phil

    Wow, those are huge stars and with only 3.8 days as their orbital period. That’s immense.

    PS. I think the “Very Large Telescope” is an awesome name.

  38. Ken G

    Tim G:
    Your estimate is probably not way off, but the power law you mention breaks down when you near the Eddington limit. At that point, the luminosity is more nearly proportional to mass to the first power– expressly to avoid violating that limit. So a better estimate is obtained by just putting this star at the Eddington limit, and that’s a solar luminosity of more like 5 million, IIRC.

  39. robmoff2

    I am no astronomer, just a kid who loves space and forgot to grow up, so I can’t offer any insight or technical background information. That being said… WOW!

    The masses and distances and orbits involved here are mind blowing! Then with the discussion about 2 galaxies colliding without any impact… my head hurts.

  40. whoa, that is a really big STAR, you know who else is a BIG star? ..well rap superstar Mike Jones.

  41. tjm220

    Nice spam Mike Jones, I’m under whelmed by the segue.

  42. Ryan

    “the centrifugal force would be ferocious, I’d wager, and the upper layers of those stars are barely holding on due to the Eddington limit.”

    There is no such thing as “centrifugal force” There is only centripetal force, or acceleration + inertia.

  43. Ken G

    There is no such thing as the force of gravity either– yet they are both very useful constructions in the hands of someone who knows how to use them. Indeed, that’s a pretty good definition of “physics”.

  44. You think Very Large Telescope (VLT) sucks ??? What about installing a smaller telescope (2,6 m) next to the big four, and call it the VST ??? The real meaning is VLT Survey Telescope, but everybody here (I work at Paranal) calls it the Very Small Telescope.

    The smaller enclosure to the right.

  45. John Castner

    Alex, that’s gonna make me grin all day. A 2.6 meter scope known as the Very Small Telescope. Isn’t that ironic (at least in an Alanis Morrisette sense)?

  46. Jean-Claude

    Thank you so much for such a wonderful, enlightening explanation of the findings.

  47. Grand Lunar

    Excellent entry, Phil! Nice to read REAL science admist all the creation nonsense.

    I’m also enlightened by the insight about star masses. I didn’t know that Eta Car’s mass wasn’t actually confirmed like this star was. The same holds true for the Pistol star, I assume?

    Well, even 114 solar masses is BIG.

  48. SCR

    A bit late now but in case people check later …

    Frank Said : [On June 7th, 2007 at 8:07 am]
    “Any real possibility that these two stars could collide someday?”

    If their orbits decay quickly enough they may actually _merge_ forming a peanut shaped contact binary like W Ursae majoris or ER Vulpeculae forming an FK Comae Berenice type star with the combined masses of both stars … !!

    So 114 + 84 = A single 198 mass hypergiant — WHOOOAAAA!!!!

    That ‘s assuming the fusion of the twain doesn’t then blow itself apart having grossly exceeded the Eddington limit … as would Isuppose probabaly happen -creating some sort of huge nebula or perhaps a supernova .. ??? Any modelling done yet on these stars evolution over time … ??

    Eta Carinae, incidentally, was listed by James B. Kaler, a stellar expert (author, Professor & sometimes ‘Astronomy’ magazine contributer)
    as a binary with a 60 solarmass Wolf-Rayet star & an 80 solar mass B0 blue hypergiant star orbiting each other every 5.6 years.

    So ‘Macz’ & y’all, Stellar mergers or collisions can and do happen – with the provisio that this happens in binary and multiple stellar systems.

    BTW. These must be astonishing super-luminous stars – any idea of th absolute Magnitudes (visual & bolometric) and spectral types, folks?

    Finally, is it true that there are no stars “earlier” than type O3 (ie no O0, O1, O2 stars) known? Could these be the first examples ever found?

  49. Richard

    That’s a pretty massive star, but if 114 solar masses is the greatest mass known to this date, how many times the Sun’s mass does LBV 1806-20 have?

  50. will

    that was the BIGEST STAR I EVER SAW

    [-_-][-_-]:-] =-]

  51. Mitch in Toronto

    centrifugal forces? BAD SCIENCE! There is no string attaching the two orbiting masses together. You cannot cut anything and have them fling apart. Holy Crow – who’s this coming from? Mr. Bad Astronomy himself. Ai Karamba!


Discover's Newsletter

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


See More

Collapse bottom bar

Login to your Account

E-mail address:
Remember me
Forgot your password?
No problem. Click here to have it e-mailed to you.

Not Registered Yet?

Register now for FREE. Registration only takes a few minutes to complete. Register now »