≡ Menu

Stars at Galactic Escape Velocity

How do you boost the velocity of a star up to 540 kilometers per second? Somewhere in that region, with a generous error range on either side, is the speed it would take to escape the galaxy if you left from our Solar System’s current position. Here on Centauri Dreams we often discuss exotic technologies that could propel future vehicles, but it’s hard to imagine mechanisms that would drive natural objects out of the galaxy at such speeds. Even so, there are ways, as explained by Vanderbilt University’s Kelly Holley-Bockelmann:

“It’s very hard to kick a star out of the galaxy. The most commonly accepted mechanism for doing so involves interacting with the supermassive black hole at the galactic core. That means when you trace the star back to its birthplace, it comes from the center of our galaxy.”

The mechanism works like this: A binary pair of stars moving a bit too close to the massive black hole at the center of the Milky Way loses one star to the black hole while flinging the other outward at high velocity. When you calculate that the black hole has a mass equal to some four million Suns, this works: Stars can indeed be accelerated to galactic escape velocity, and so far a number of blue hypervelocity stars have been found that could be explained this way.

But Holley-Bockelmann and grad student Lauren Palladino have run into something that casts doubt on this explanation, or at least makes us wonder about other methods for making stars travel this fast. Calculating stellar orbits with data from the Sloan Digital Sky Survey, the duo have found about twenty stars the size of the Sun that appear to be hypervelocity stars. Moreover, these are stars whose composition mirrors normal disk stars, leading the researchers to believe they were not formed in the galaxy’s central bulge or its halo.


Image: Top and side views of the Milky Way galaxy show the location of four of the new class of hypervelocity stars. These are sun-like stars that are moving at speeds of more than a million miles per hour relative to the galaxy: fast enough to escape its gravitational grasp. The general directions from which the stars have come are shown by the colored bands. (Graphic design by Julie Turner, Vanderbilt University. Top view courtesy of the National Aeronautics and Space Administration. Side view courtesy of the European Southern Observatory).

Holley-Bockelmann and Palladino are working on possible causes for the movement of these stars, including interaction with globular clusters, dwarf galaxies or even supernovae in the galactic disk. The list of possibilities is surprisingly long, as noted in the paper on this work (internal references omitted for brevity):

While the SMBH [supermassive black hole] at the Galactic center remains the most promising culprit in generating HSVs [hypervelocity stars], other hypervelocity ejection scenarios are possible, such as a close encounter of a single star with a binary black hole… In this case, the star gains energy from the binary black hole and is flung out of the Galaxy while the orbit of the black hole binary shrinks…Another alternative hypervelocity ejection model involves the disruption of a stellar binary in the Galactic disk; here a supernova explosion in the more massive component can accelerate the companion to hypervelocities…

We may know more soon, for the paper points out that a nearby supernova should have contaminated the spectrum of a hypervelocity star. As they delve into these and other possibilities, the researchers are also expanding their search for hypervelocity stars to a larger sample within the Sloan data to include all spectral types.

The findings were announced at the meeting of the American Astronomical Society in Washington this week. The paper is Palladino et al., “Hypervelocity Star Candidates in the SEGUE G and K Dwarf Sample,” The Astrophysical Journal Vol. 780, No. 1 (2014), with abstract and preprint available.


Comments on this entry are closed.

  • Moebius January 10, 2014, 11:06

    Even if it’s more likely that they are accelerated by a natural event, wouldn’t it be possible that we are actually seeing stars artificially accelerated, for instance by a “half- or quarter-Dyson-sphere” sail, like in Bowl of Heaven?

  • CharlesJQuarra January 10, 2014, 11:21

    Interesting places to look for Shkadov thrusters

  • Ron S January 10, 2014, 12:12

    I had to wonder if any of these HVS might have a source in another galaxy, dwarf or otherwise. The authors address this point in the paper (it’s just 8 pages, so a quick read). After all, if our galaxy can generate HVS so can others. We are not exceptional. As with stars exchanging Oort cloud objects galaxies could “exchange” HSV.

    They estimate 1 Gy to get here from M31, which a quick estimate of mine shows to be about right for the velocities observed. So the stars have to be right class to survive the journey. O and B stars don’t make it. They find G and K stars in their survey, but none has the right trajectory to come from M31. They cannot exclude other possible extra-galactic sources. But to come from further afield excludes any other than long-lived stars, and most of those would not intersect another galaxy within their lifetimes.

    So purely on a statistical basis the HSV very likely have a galactic origin and, due to their trajectories, cannot have the galactic SMBH as a source.


  • James D. Stilwell January 10, 2014, 13:47

    This post brings to mind a related unsettling matter….
    Don’t discount Halton Arp’s theory that quasars might be kicked out of a galaxy (from our point of view) with a very high red shift and therefore might not be the distant objects they’re currently believed to be….
    Remember when neutrinos were once thought to be massless…..
    ctp Berkeley Neutrino Physics still believes they have mass….
    But this is another can of worms…..Time marches on doesn’t it….
    So many balls to keep in the air….I get dizzy reading your posts….

  • Lepton January 10, 2014, 15:58

    If our Sun was flung out of the MW, and we developed on Earth after that, we will not see a single star in our night sky, we only see planets, moons and galaxies, that would be quite a different story for our effort to understand the universe.

  • Michael January 10, 2014, 16:06

    Perhaps some of these hypervelocity stars are the result of close binaries in which one of the Stars has gone supernova, type 1A (total destruction), stealing gas from the other. If the right star at the right age is present in the system it could allow it to get quite close to the soon to be supernova star creating high orbital ejection velocities. There would most likely be a signature of elements like carbon and iron excesses in their atmospheres.

  • coolstar January 10, 2014, 18:27

    @Moebius Bowl of Heaven is the most frustrating novel I tried to read over the holidays (I finally gave it up as a bad job). The humans never even mention that the object they come upon is moving at many, many times the galactic escape velocity (which would be the VERY first thing to notice, well before it’s resolved)! (or importantly, that it has to undergo a constant acceleration and cannot either slow down OR change direction). It’s a disgrace and reads like a rough (very) draft. They even get the spectral type of the star very, very wrong. A list of all the physics and plotting errors would go on for many pages. (one can find such lists in the Amazon reviews, if one is that much of a masochist). End of mini-rant.
    Back to real run-away stars: I’ve always assumed that SN were the main source of such stars. Just from the statistics of core collapse SN I would, off the top of my head, guess that they are much more likely than interactions with Sgr A*.

  • Michael Spencer January 11, 2014, 8:37

    What coolstar said. I thought the same thing even though Jack McDevitt recommended the novel…to me, it read like Niven had left over Ring material.

    Some of the other ideas were cool, though.

  • Brett Bellmore January 11, 2014, 12:40

    Does a star have to get it’s escape velocity in one event, rather than a fortuitous sequence of encounters? What I’m saying is, do these stars represent a separate peak in the distribution of star velocities, or are the just the tail of a thermal distribution?

  • Michael January 11, 2014, 16:18

    These stars may also get a kick from the supernova explosion.


    A ‘Jupiter’s worth of material blown off the companion star by the titanic blast.’ travelling at what looks like 4700 ± 100 km s–1


    And another article about it


  • Gregory Benford January 11, 2014, 16:25

    Coolstar: Your points are dealt with in the sequel, SHIPSTAR, due out April; not all is as it seems.

    So do the HVS velocities point back to the center of the galaxy?

  • ljk January 11, 2014, 18:59

    In the Star Trek novel Dyson Sphere, the written sequel to the 1992 ST:TNG episode “Relics”, an advanced race has flung a neutron star at the Dyson Sphere as part of some ancient interstellar conflict.


    Since we are having so much fun speculating, maybe that is what happened to those stars: Ammunition from some hitech space weapon or the collateral damage from an interstellar war.

  • Michael January 11, 2014, 21:59

    @Gregory Benford January 11, 2014 at 16:25

    ‘So do the HVS velocities point back to the center of the galaxy?’

    frustratingly not!


    There is a possibility that a more massive star is orbited by a highly eccentric smaller star which makes close approaches. As the larger star goes in the red giant phase the smaller star will collect its gas and get even closer to the star due to tidal forces. If close enough the white dwarf that has formed from the more massive star will collect the gas off the orbiting star near closest approaches and gain mass up to the Chandrasekhar limit. When that limit is reached the star destroys itself in a supernova allowing the orbiting star to be freed. Now the release point of the star is more likely when the orbiting star is at closest approach due to the release of more mass building gases from the orbiting star, the result could be a very high velocity ejection.

  • Rob Henry January 12, 2014, 5:13

    As I have said before Shkadov thrusters work perfectly to get to these sort of velocities if they are used purely for course adjustment to harness the slingshot effect. So what if they are so weak that that adjustment must be made two previously encounters previously to the current one for ultimate effect. The computing power needed to work that out is not all that high.

    And why would an ETI want to get to such a velocity while contained within a disk? To optimise the number of new systems they can colinise by close encounter in a given time!

    And Brett Bellamore, if any such mechanism worked by accidental means we would see thousands of stars at three quarters this excess velocity, and millions and half it.

  • A. A. Jackson January 13, 2014, 12:24

    This brings back memories.
    When I was a grad student at UT in Austin my graduate mechanics class was taught by stellar evolution astrophysicist David Arnett. One our first problems was to figure out in a binary system with a massive star about to go supernova how much mass would the supernova have to loose (to infinity) for the system to become unbound? I was kind of stumped. I little library work showed me this was the ‘Blue Runaway’ problem, actually known now as the OB Runaways. They were discovered by the Dutch astronomer Adriaan Blaauw in the 1950’s. A little later Blaauw even presented a model solution, the supernova had to lose one half it’s mass. Turns out not a hard calculation to do … but I needed a gouge , like that, to complete it!
    Later I made friends with Los Alamos physicist Jack Hills, a remarkable man , he did a lot of things. In the 80’s he thought that if one scattered binary stars off massive black holes one would be captured by the hole and one ejected to infinity. Numerical simulations showed this was the case.
    (Actually the black holes don’t have to be super massive, globular clusters may do the same thing.)
    Turns out Blue Runnaways can be made the same way … but seems both methods work.
    Some controversy about the supernova mechanism but that just seems a like it’s still true.
    So the hypervelocity star production was predicted before they were observed!
    Note 1000 km/sec is .003 the speed of light, not bad… but only good for STL travel.

  • David January 14, 2014, 1:00

    Brett Bellmore,

    The separation between stars is so great that these velocities are almost certainly
    from single events. Furthermore, the stellar velocity distribution in our galaxy
    is far from being thermal. The relaxation time, which is the time required for a thermal distribution to arise is about three orders of magnitude higher than the current age of the Universe.

  • Michael January 14, 2014, 14:51

    Using Keplers laws it can be seen that a white dwarf at about the critical mass of 1.44 solar masses can after going supernova (type 1A) eject a 1 solar mass star out at a very high velocity if the explosion occurred at periastron.

    For example using a system eccentricity of 0.9 and a distance of 700000km between stars at closest approach yields an ejection velocity of approx. 1700 km/s.
    Doubling the distance between the stars and keeping the eccentricity at 0.9 yields an ejection velocity of approx. 1200 km/s. It appears that the equation is more sensitive to eccentricity than distance between the stars.

    So it is feasible on paper!

  • coolstar January 16, 2014, 18:04

    Sorry Dr. Benford, but as a life long fan of your writing (I can’t remember the number of times I’ve recommended TIMESCAPE as the finest fictional example of how science is really done), I won’t be reading SHIPSTAR.

  • coolstar January 16, 2014, 18:08

    @ A. A. Jackson. Yep, I did the same calculation in a grad astrophysics class. but as you alluded to, it’s really a first or second year undergrad problem (especially after one knows the answer!).

  • ljk January 23, 2014, 18:19

    There is a supernova in Messier 82. I wonder if it flung any stars out of that galaxy at hypervelocities?


  • Michael January 24, 2014, 13:49

    @ljk January 23, 2014 at 18:19

    ‘There is a supernova in Messier 82. I wonder if it flung any stars out of that galaxy at hypervelocities?’

    Possible, but I doubt very much observable, the blast would produce an overwhelming glare blotting it out.

  • ljk May 2, 2014, 9:11

    Megagalaxy Messier 87 just flung an entire star cluster right into the intergalactic void:


    Have to wonder if anyone is “aboard” that “vessel”?

  • ljk May 2, 2014, 9:26

    The discovery on an entire star cluster being ejected from M87 made me think of this piece from Orion’s Arm: