The radiant beauty of cosmic collisions

By Phil Plait | October 13, 2009 7:00 am

What happens when two massive spiral galaxies — each with a hundred billion stars — slam into each other head-on at hundreds of kilometers per second?



[Click to embiggen; or go here to get the massive 15Mb TIF image.]

This is the unusual galaxy NGC 2623, seen in this newly-released and breathtaking Hubble Space Telescope image taken in 2007. We’re seeing this vast smash-up caught in the act, a galactic collision already in progress. It appears frozen in time, but that is an illusion of distance: at a distance of a staggering 250 million light years the tremendous velocities of the collision are reduced to a motionless tableau on the human timescale.

But we see a large number of galactic collisions when we catalog the sky, and together with our knowledge of math and physics we have a good understanding of how these encounters play out.

When two massive galaxies approach each other, the gravity of each starts to affect the other. Call them Galaxy A and B. The side of Galaxy B closer to Galaxy A feels more gravity from it, so stars and gas are drawn toward it more strongly than the stars and gas on the far side of Galaxy B. The same is true in the other galaxy. As they get closer, this force strengthens, teasing out long ribbons of material — called tidal tails — that stretch in the direction of the other galaxy.

If the encounter is off-center, then the tails get curved when the galaxies pass, arcing either gently or severely depending on the speed, encounter distance, and mass of each participant. The Hubble image clearly shows the arcing tails from each galaxy in NGC 2623.

Incredibly, even though hundreds of billions of stars are involved, each individual star is far too small to suffer a physical collision. But gas and dust clouds are much bigger than stars (they can be hundreds of trillions of kilometers across, as opposed to stars which are a trifling million or so kilometers in diameter), so collisions between them are common. When clouds collide they collapse and undergo violent bouts of star formation. This too is clear in the image: the blue clumps in the tidal tails are vast regions of clusters of stars being born; over 100 such clusters have been identified in this image in the tail on the right alone.

Collisions like this blast out energy, not just in visible light, but at other wavelengths as well. In infrared alone, NGC 2623 radiates with the power of 400 billion times the Sun’s energy. This makes NGC 2623 a ULIRG: an ultraluminous infrared galaxy. Although relatively rare locally, they are so common at great distance (and therefore earlier on in the age of the Universe) that they comprise as much as half of all the infrared background glow we see in the Universe. The huge amount of infrared comes from the collision itself; star formation produces prodigious amounts of dust which absorb ultraviolet light from newly-born stars and re-radiate it in the infrared. The collision also dumps gas and dust into the central supermassive black holes in the cores of the two colliding galaxies, which piles up in a flat disk outside the black hole, heats up hugely, and again glows brightly.

Astronomers are making a comprehensive study of such ULIRGs using a fleet of telescopes including Hubble, Spitzer, Chandra, GALEX, 2MASS, VLA, and even the venerable IRAS satellite which surveyed the sky in infrared in the 1980s, and in fact first discovered the ULIRGs.

Why study them? Because galaxies as large as our Milky Way almost certainly started off small and grew to their present size by colliding and merging with other galaxies. Studying ULIRGs is a way of examining how our galaxy came to be… and it’s a glimpse of our future as well. In a billion years or more, we will suffer a massive collision with the Andromeda Galaxy. Our own clouds of gas and dust may smash into those in Andromeda, creating huge waves of star formation and blasting out light at all wavelengths. What will our fate be then? The Earth may survive — the Sun will still be around for this event — and the gravitational repercussions may toss us out of the new galaxy, or drop us down to the core.

It may seem academic, but astronomers thirst for understanding of these events. We want to know how we came to be, and where we are headed. That knowledge may have little or no practical use for our own survival (or at least not for a few million millennia), but for now, for today, we learn more about galaxies in general, more about the physics of cosmic collisions, and more about the interaction of gas and dust on a truly mind-numbing scale.

And of course, we get to gaze on lovely images, illusions of placidity and gentleness to be sure, but lovely nonetheless.

Image credit: NASA, ESA and A. Evans (Stony Brook University, New York & National Radio Astronomy Observatory, Charlottesville, USA)

CATEGORIZED UNDER: Astronomy, Pretty pictures

Comments (52)

  1. And what will the night sky look like on Earth, when Andromeda is close enough to fill our field of view?


    Hey Phil, at the ninth paragragh, in the fifth line, I think that term should be comprise (to consist of; be composed of; to constitute; to include; contain), not “compromise” (a settlement of differences; a concession).

  3. Ah, thanks IVAN3MAN. Must’ve hit the wrong link on the spell checker. :)

  4. LibraryGuy

    Quick question: Have astronomers ever seen something that might be the result of two stars running into each other?


    @ LibraryGuy,

    Blue straggler stars are hypothesized to be current or former binary stars that are in the process of merging or have already done so. The merger of two stars would create a single star with larger mass, making it hotter and more luminous than stars of similar age.

  6. Peter B

    Would a galactic collision create conditions hostile to life? That is, would the radiation be that serious? Or are the effects fairly localised/minor?

  7. nobody

    @ Phil:
    I am the nobody guy from the “Mr. Deity gets squishy” post #32. Posts like this one is what I expect from this blog! Keep up the good work 😉

  8. Why do you have to call them Galaxy A and Galaxy B? Can’t you call them Galaxy 1 and Galaxy A? Or does one of them feel like a “B?”


  9. Ari

    Out of curiosity, I fired up google sky after seeing the picture:

    “wow,” I said to myself, it’s also pretty bright in infrared :)
    (you have to zoom a level or two to get the galaxy in visual)

  10. gwmsfl

    embiggen?? How much crediblity can a website have when they make up their own words?

  11. sophia8

    I’d like to see that kind of galaxy-smashing close up. Not too close-up, mind – a system on the trailing end of one of the ‘tails’ would be far enough away from the radiation glare while still getting a spectacular nighttime sky.

  12. Paula

    Bravo Phil Plait!
    I’m finishing my phd on galaxy mergers and it’s really, really nice to read something so… kind about something so complex! :)

  13. I’m glad you approve nobody. 😉

    Peter B, we really aren’t 100% sure, but since galaxy collisions are mostly empty space hitting more empty space, there really isn’t a lot going on during the merging. Of course, some regions (areas with a lot of interstellar dust and material) may cause problems, but we don’t have a lot of info to go on as far as I am aware. Also, the merging of the central black holes could be spectacular. (This is just my layman understanding, you’d need to talk to a real astronomer… wonder where we could find one of those?) 😛

    EDIT TO ADD: Paula, since I have a restraining order from harassing Dr. Plait, mind answering Peter B.’s question? My layman is probably very weak!

  14. DaveS


    I’m curious, though, how do astronomers know that the blue area is an area of star formation? Surely the time scale is way too long to observe directly.

  15. @gwmsfl embiggen isn’t made up. It’s a perfectly cromulent word.

  16. ZB

    @gwmsfl re. “embiggen” : It’s a perfectly cromulent word.

  17. Gorgeous photo. I love these posts as much as I love looking upward.

    BTW @gwmsfl, “embiggen” is just a fun creative word. This site is far more credible and legit than most out there. Breathe and enjoy.

  18. Chris

    That’s a nice binary star in the upper left corner.

  19. SionH

    Of course, since we’re not sure of the effect of such collisions on any extant life in these galaxies, we may be looking at a cosmic Hindenburg disaster, Oh the humanity! Alienity?

    @ZB, you beat me to it :)

  20. Gamercow

    I believe many data points are observed to form a trend, and conclusions are drawn from that. That is, they see it happening, but not in one object, but rather several objects that are at different points in the time line.

  21. Ari

    DaveS: Blue stars tend to have relatively short lifetimes (several million years). Because of that, they don’t really have time to wander far from where they’re born. So if you see lots of blue stars in one place, astronomers can assume that that is a site of star formation.

  22. LibraryGuy

    @IVAN3MAN AT LARGE thanks for the info.

  23. mike burkhart


  24. brian

    The researchers are also interested in the end states of these and trying to see if they end up being elliptical galaxies. The star formation rates are huge in these merging states, and over the course of the merging process, it might be possible to exhaust the star forming gas and resemble an elliptical galaxy (stars forming at nearly the same time and having relatively little molecular gas).

  25. gbraver

    The two bright stars in the top left corner look interesting…


    LibraryGuy, you’re welcome!

    D’OH! I just noticed a typo at my post (#2) above; it should be paragraph, not f***ing “paragragh”!

  27. TaoMacGuy

    A question Phil,

    So do stars ever physically collide during galaxy collisions? And if so, what would that even be like? I have this image of Shoemaker-Levy but on a way more massificated scale!

    I suppose I’d never really thought of two stars physically colliding before and now my mind is trying to play out what sort of awesome that might be like.

  28. I thought this was an antiantivax website I did nor know it had anything to do with astronomy.

  29. Sarah

    #1, KenB …

    “And what will the night sky look like on Earth, when Andromeda is close enough to fill our field of view? ”

    Like in here?

  30. JupiterIsBig
  31. Get real

    Straight from Nasa’s Hubble Site on how the manipulate pictures. ” To depict how an object might look to us if our eyes were as powerful as Hubble”….. Um might, really?! All NASA does is try to drum up public interest with these pics. On the real side a scientist and artist collaborate on some not all but some pictures. They read out the numbers on a piece of paper about the gases and x-rays detected and paint a pretty picture. Most objects in space they admit emit to much light to take a picture of. (See also some awesome pics of black holes) Here is a quote from that page. “This is a still image from a computer artist’s animation of hot iron gas riding upon a wave in spacetime around a black hole.” Cowabunga dude!

  32. Johnfruh


    Love your post. beautifully elucidated.

    However, I do have a nit to pick, even though it is tangential to your post and it is this…

    In your 2rd paragraph (after the image), you mention the mutual gravity felt by the two galaxies. That’s fine as far as it goes.
    However, in the last sentence, you say: “As they get closer, this force strengthens,…”, (referring to gravity).
    Now then, here is my point/nit. We are now some 90+ years into the age of Einstein’s General Theory of Relativity which states that gravity is a warping of spacetime and as such not a force as Newton thought.

    So, in the inimitable words of Dezi Arnez, who oft times said,” ‘splain it to me Lucy!”, I ask, how come you are referring to gravity as a “force”? What am I missing here? Which is it? A force or a warp in the fabric of spacetime?

    I find these sorts of “cavalier” references to gravity everywhere and it bugs the hell out of me. So, please, seriously, tell me what is going on. Is it just sloppiness of terminology? Is it that you are just falling back on “common usage” when calling gravity a “force”? Am I being just too literal/anal? Or what?



  33. Buzz Parsec

    #28 & #31,

    ISTR that direct star collisions are extremely unlikely but not impossible. Near misses (which are also called collisions, when the stars come close enough to each other that their gravity drastically alters their orbits around their galaxies and disrupts their planetary systems, an unfortunate choice of terminology IMHO) are much more likely, and can send stars flying off into intergalactic space or hurtling towards certain doom (in a few million years) by falling into one of the black holes at the centers of the colliding galaxies.

    Someone should write a book about this. It could be a general book about all the horrible things that could happen to the Earth or the Sun or the solar system, and it could be discussed in the chapter on bad stuff our galaxy could do to use, on page 258.

    BTW, did anyone ask either of the galaxies if they minded being called Galaxy A and Galaxy B? I find this very derogatory to Galaxy B, and even if it said it was okay to call it that, how do we know it wasn’t pressured into it, just to be accepted? People (and galaxies) will put up with all kinds of abusive behavior just to be accepted. This doesn’t mean we should allow clearly disfunctional galaxy families to persist unchallenged. It is clearly time for an intervention.

  34. Buzz Parsec


    Contrary to popular belief, Einstein’s theories did not invalidate either Newtonian mechanics nor Newtonian gravity. For almost all the universe and almost everything in it, almost all the time, Newton’s view is just fine. Except in very rare instances such as in black holes and gravitational lenses or shortly after the Big Bang, or over extremely long time spans (10’s of billions of years), General Relativity produces the same results or only a very trivial correction to Newtonian gravity, which is mathematically much simpler. Physicists use Newton’s theory all the time for most problems, complete with forces and vectors and point masses and so forth. The key is knowing when it is an appropriate approximation and when it is not.

    For example, when sending a spacecraft to Mars, NASA uses Newtonian physics to predict its trajectory and to compute course corrections. This works just fine (except when they screw up the units) and the calculations are much more accurate than the measurement errors (of the spacecrafts position and orientation) and the throttling and timing control of the engines. There is no need to apply general or special relativity to the problem, and even if you bothered to do so, all it would tell you is you should run the engine at full thrust for 4 minutes 37.250073 seconds instead of for 4 minutes 37.25 seconds for an optimal orbital insertion at the desired distance from Mars, for example.
    (Numbers made up for illustrative purposes…)

  35. Johnfruh

    Thanks Buzz Parsec.

    I do understand that Newton’s concepts and laws of gravity are a good approximation to what really is going on. And, yes, at our pedestrian pace, it works well enough, its inability to account for the precession of Mercury’s perihelion notwithstanding.
    However, note, that if we used Newton’s approximations alone, the GPS system would not work!

    But that is not my point. In this discussion, I’m not so much interested in the practicalities of the use of either Newtonian v. Einsteinian math to work out orbits, trajectories, etc. What I’m asking is whether or not there IS such a thing as a “force” of gravity or are we just using that term as a convenient shorthand? Even Newton was only able to describe the effects of gravity with his famous equation. However, he still had no idea what it WAS/IS (if I remember correctly).

    So again, I ask, what IS it? A warp in spacetime or a force?
    And, when Phil is talking about Galaxies feeling the mutually “attractive force” of gravity between them, would it not be more accurate to talk about the spacetime warp experienced by the different far flung regions of each galaxy.
    As a layman, I’m just wondering, that’s all.
    But, playing devil’s advocate for a moment, can’t you just see nay-sayers using this loosy-goosy way of referring to gravity as a weakness (i.e. being charged with talking with a forked tongue)? Or am I being, again, too anal?


  36. Petrolonfire

    @ 32. Johnfruh Says:

    Phil, Love your post. beautifully elucidated

    Huh? Elucidated? Is that another made up but perfectly cromulent word like “embiggen”? 😉

    Now “embiggen” I get but “elucidated” – what doe’s that mean?

    Great photo! :-)

  37. Messier Tidy-Upper

    Fantastic image – thanks BA! 😀

    @ 1. Ken B Says:

    And what will the night sky look like on Earth, when Andromeda is close enough to fill our field of view?

    If I recall right from reading a certain familiar book (think it had the word ‘death’ in its title, maybe ‘Skies’ as well, ‘from’ and ‘the’ were in there somewhere too I think! 😉 ) won’t this Galactic collison between the Milky Way and Andromeda take place at a time when our Sun has evolved into a red giant star?

    Also won’t this mean our future scorched Earth then becomes tidally locked so that one side faces the Sun and t’other away permanently?

    If so, then your view will depend on where you are.

    On the (very!) Sunny side of Earth, you won’t get much of a view of the Great Galactic Collision – but you will see a humungous gargantuan blood red Sun taking up much of the starless black sky and not-so co-incidentally, also get fried by immense amounts of radiation, light and heat and probably vanish into a molten lava sea. Think Mustafar from ‘Star Wars : Ep. III Revenge of the Sith’ – only with less solid land! 😉

    On the Dark Side of the Earth though, you’ll have a superb view of Andromeda as it approaches and of the tidal tails of stars drawn out from both galaxies and the starburst phase will fill the sky with millions of bright massive stars like the Plieades – and these stars will also be in rich star clusters with lots of nebulas and filaments of colliding and ejected gas clouds and all sorts like that.. Yegods, it’ll be beautiful! But also cold. Very, *very* cold. Think Hoth from Star Wars : Ep. V The Empire Strikes Back’ only darker and with no air*, daylight or tauntauns. So I guess more the farside of the Moon or Triton or Europa really! 😉

    (Starts reconsidering the idea that every single planet possibly imaginable can be best visualised using analogies to the various planets in ‘Star Wars.’ 😉 )


    * No air beacause I imagine it will all either freeze or be swept away and lost into space because of the intenser solar radiation and increased solar
    winds. Right?

  38. “Embiggen” has 76,000 results in Google.

    So far, astronomers have found some 200-300 planets. They estimate 5% of stars have planets. Galaxy A + B each have a billion stars, or some 50 million planets each. There may well a Phil Plait on one of those planets who is writing about it for his fellow sentient beings.


  39. Johnfruh

    @ 36 PetrolonFire

    Elucidate: to make lucid or clear; throw light upon; explain:


  40. Buzz Parsec

    @ 36 Johnfruh,

    I think it depends on what model you used to do your calculations… If you use Newton’s laws, you talk about forces. If you use GR, you describe the situation in terms of warped spacetime.

    The point I was trying to make, though, wasn’t really about gravity, forces, or colliding galaxies. It was about how science works… Lots of woo purveyors claim that if science is provisional and subject to change, and has in fact changed many times (such as the replacement of Newton’s theory of gravity by Einstein’s), then what it says today isn’t what it is going to say tomorrow, so why should be pay any attention to it? They are ignoring the requirement that any new theory has to accurately describe existing data. For example, for GR to replace Newton’s theory, it has to reduce to Newton’s theory in the cases where we know Newton’s theory works extremely well, and any replacement for GR (e.g. a successful quantum gravity theory) has to reduce to GR in cases where we know GR works, and to Newtonian gravity in cases where we know that works. What’s more, the theory has to include some way to decide when earlier theories suffice, it can’t just be arbitrary. (E.G. “On Tuesdays, the force of gravity points up and is an inverse cube law instead of an inverse square law, because God needed it that way to create mud.”)

  41. Buzz Parsec

    P.s. and thanks for elucidating “elucidate” :-)

  42. Petrolonfire

    @ 42. Johnfruh Says:

    @ 36 PetrolonFire Definition: Elucidate: to make lucid or clear; throw light upon; explain: …John

    Thanks. Ditto to #44 Buzz Parsec & seconded there.

    So .. was that a “real” word all along then? (Blushes.)

  43. johnfruh

    Buzz @ 43

    Thanks again, Buzz. I do appreciate that the two models can be used to describe gravity and have their own uses in various circumstances.

    I believe that I do understand your point and have no quarrel with it. Thanks also for the bit about the woo purveyors. Well said.

    However, I don’t think that I have made my point clear as I still do not have an answer for it (although it might be hinted at by your use of the term “Model”).

    So, let me try again. IS gravity a force or a curvature of spacetime? Or are both concepts just a “model” of gravity? If so, are we still no closer to understanding what it IS?

    Tangentially, would it not be better to describe the effect of gravity on galaxies using GR terminology? After all, as you said, GR applies at large scales and galaxies are rather large. For instance, could it be that the curvature of the galactic tails is somehow due to the time it takes for the gravity at the centre of the interacting galaxies to get out there (after all, gravity can only “travel” at the speed of light, right?

    Petrolonfire @47 … You’re welcome. And, no, embiggen is not a real word. I think that Phil is playing with a take off from the real word “enlarge”.



  44. Earth sci guy

    Brilliant photo…great, colorful article…and the responses here are like watching Big Bang! There is nothing wrong with using/creating words like ’embiggen’. It adds a bit of spice to life. Relax, enjoy, laugh!

  45. Just so we’re all clear, Embiggen (and Cromulent) are original words coined for/by the Simpsons cartoon TV Series, see for an explanation :)
    It’s all a just a bit of fun really!


Discover's Newsletter

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


See More

Collapse bottom bar