Star-Birth Myth Shattered


An international team of astronomers has debunked a long-held belief about how stars are formed.

Since the 1950’s, astronomers believed groups of new-born stars obeyed the same rules of star formation, which meant the ratio of massive stars to lighter stars was pretty much the same from galaxy to galaxy.  For every star 20 times more massive than the Sun or larger, for example, there’d be 500 stars equal to or less than the mass of the Sun.

“This was a really useful idea. Unfortunately it seems not to be true,” said team research leader Dr. Gerhardt Meurer of Johns Hopkins University in Baltimore.

This mass distribution of newly-born stars is called the ‘initial mass function’, or IMF.  Most of the light we see from galaxies comes from the highest mass stars, while the total mass in stars is dominated by the lower mass stars which can’t be seen, so the IMF has implications in accurately determining the mass of galaxies.  By measuring the amount of light from a population of stars, and making some corrections for the stars’ ages, astronomers can use the IMF to estimate the total mass of that population of stars.

Results for different galaxies can be compared only if the IMF is the same everywhere, but Dr. Meurer’s team has shown this ratio of high-mass to low-mass newborn stars differs between galaxies.  Small ‘dwarf’ galaxies, for instance, form many more low-mass stars than expected.

To arrive at this finding, Dr. Meurer’s team used galaxies in the HIPASS Survey (HI Parkes All Sky Survey) done with the Parkes radio telescope near Sydney, Australia.  A radio survey was used because galaxies contain substantial amounts of neutral hydrogen gas, the raw material for forming stars, and the neutral hydrogen emits radio waves.

The team measured two tracers of star formation, ultraviolet and H-alpha emissions, in 103 of the survey galaxies using NASA’s GALEX satellite and the 1.5-m CTIO optical telescope in Chile.

Selecting galaxies on the basis of their neutral hydrogen gave a sample of galaxies of many different shapes and sizes, unbiased by their star formation history.

H-alpha emission traces the presence of very massive stars called O stars, the birth of a star with a mass more than 20 times that of the Sun.

The UV emission, traces both O stars and the less massive B stars — overall, stars more than three times the mass of the Sun.

Meurer’s team found the ratio of H-alpha to UV emission varied from galaxy to galaxy, implying the IMF also did, at least at its upper end.

“This is complicated work, and we’ve necessarily had to take into account many factors that affect the ratio of H-alpha to UV emission, such as the fact that B stars live much longer than O stars,” Dr. Meurer said.

Dr. Meurer’s team suggests the IMF seems to be sensitive to the physical conditions of the star-forming region, particularly gas pressure.  For instance, massive stars are most likely to form in high-pressure environments such as tightly bound star clusters.

The team’s results allow a better understanding of other recently observed phenomena that have been puzzling astronomers, such as variation of the ratio of H-alpha to ultraviolet light as a function of radius within some galaxies.  This now makes sense as the stellar mix varies as the pressure drops with radius, just like the pressure varies with altitude on the Earth.

The work confirms tentative suggestions made first by Veronique Buat and collaborators in France in 1987, and then a more substantial study last year by Eric Hoversteen and Karl Glazebrook working out of Johns Hopkins and Swinburne Universities that suggested the same result.

Source: CSIRO

20 Replies to “Star-Birth Myth Shattered”

  1. The results of the above mentioned research, lead by Dr. Gerhardt Meurer, can be found here:
    Evidence for a Non-Uniform Initial Mass Function in the Local Universe [PDF].

  2. Man, I’m gonna need several cups of strong coffee to get through reading that document! 😐

  3. … i had no idea that astronomers had thought this since the 1950’s and now that i know, i’m surprised that they thought it in the first place. does anyone know what observations the old assumption was based on?

  4. I think, you mean this:…121..161S

    The actual (or well known) Salpeter-IMF is in the “Discussion”-chapter.

  5. Could this be simplified? How are stars formed now? I’m having trouble pulling that tidbit out.

  6. @Kevin F.: it’s not easy to simplify without, almost certainly, creating misleading impressions (but here goes; caveat lector):

    Consider a big cloud of gas and dust.

    As it has mass (duh!), it will collapse under its own gravity … unless there is something to oppose gravity.

    One thing that’s very good at opposing gravity, for a ball of gas (or plasma) is pressure due to the gas being hot; one very good source of heat (to make, and keep, the gas hot) is nuclear fusion. Objects which don’t collapse further because their innards are hot are called ….

    As a big cloud of gas and dust collapses, it fragments; eventually the fragments become small (and dense) enough to become stars, or get exiled from the cloud by the radiation pressure of the new-born stars.

    Some of the stars born from the cloud will be quite massive (up to 100 sols, perhaps), some not so massive, and some will be runts (mass as low as 0.08 sols, perhaps).

    For each 1 sol mass star formed from the collapse of a big cloud, how many 100 sol mass stars will form (on average)? how many 10 sol mass stars? 0.1 sol mass stars? … and so on.

    In mathematical form, the answers to the above questions is the IMF (initial mass function).

    No doubt you’ll have questions; ask away!

  7. @IVAN3MAN: if you can grok it, no matter how many cups of coffee you have to drink, CONGRATULATIONS! 🙂

    To get the highlights, you might want to concentrate on the summary (section 7), and work backwards.

    Other than DrFlimmer, does anyone else who has read the original paper think they have understood it, in full?

  8. I think it’s a bit unfair to call this a ‘belief’ and a ‘myth’. The universal IMF is/was a simple idea that looked like it might be true. It’s just Occam’s razor – you believe the simplest explanation until the facts force you into something more complex.

  9. Since the 1950’s, astronomers believed groups of new-born stars obeyed the same rules of star formation, which meant the ratio of massive stars to lighter stars was pretty much the same from galaxy to galaxy. For every star 20 times more massive than the Sun or larger, for example, there’d be 500 stars equal to or less than the mass of the Sun.

    This article that quotes the above is very very misleading. The IMF has never been considered as the same from galaxy for galaxy – especially as observationally it is clear it isn’t. Many of the articles that came up with the IMF in the 1980s were considered more on the populations of the Milky Way. As there was no evidence in other galaxies for this distribution, it was only assumed to be the same to formulate useful theory.

    Results for different galaxies can be compared only if the IMF is the same everywhere.

    This has never been the case in past research, and if it were, it was made because there was no evidence either way. (I wrote a galaxy formation paper on this subject many years ago. I.e. Joseph Silk certainly commented specifically on this, but I will have to search for it.) Clearly the IMF MUST vary from galaxy to galaxy – as you have only to compare the ellipticals or irregulars to the spiral to know this is true. Worst the IMF must change over time as the galaxy evolves. I.e. In the early part of some galaxy’s existence, the number of stars (and their mass distribution) would have been very different. Furthermore such an idea would easily explain why star formation had stopped early in the history of elliptical galaxies, but formed stars at a less prodigious rate in spiral galaxies so that star formation could continue as it today – say in our Milky Way.

    I would have thought that it was the Star Formation Rate (SFR) that was the decided factor to determine the stellar outcomes. I.e. Faster the average formation, the smaller the stars will be produced. It is actually a distribution, where the probability of the range of stellar masses is based on the SFR not the IMF. For galaxies the star formation rates are obvious between galaxies – like around the spiral arms for example in spiral galaxies.

    If the actual quote by Dr. Gerhardt Meurer given here is true, it seems he has just massaged the truth somewhat to draw attention to his paper. (It is probably another classic example of scientists dumbing down the research which is further distorted by the media. If there is to be any debunking, it should begin with the media.

  10. @HSBC*: I share your unease concerning the rather substantial difference between what’s in the paper and what’s in the PR.

    OT: did you know that your moniker, abbreviated, is the same as the name of a large international bank?

  11. In the paper published by Meurer, “Evidence for a non-uniform initial mass function in the local universe.“, it is said in the 1. Introduction

    A key concept in our understanding of the evolution of galaxies is the Initial Mass Function (IMF), which gives the statistical distribution of masses of stars that form in a single event.

    The assumption here is very odd, because clearly the IMF in star formation is never constant over time of galaxies. Why, simply because the star formation process redistributes some of the material. Current theories assume over time the star formation rate goes through periods of activity then periods of quiescence. Hence high surface brightness galaxies will have lots of O-type stars, while during the quieter times, most of the massive O-type stars have disappeared as mostly supernovae.
    So whilst it may or may not be true galaxy to galaxy varies in IMF …IN A SINGLE EVENT, it certainly not the case over the whole evolution of the galaxy.

    Also interesting is the authors conceding this idea is not also new, I.e.;

    Ours is not the ?rst study to probe the nature of the IMF using H? and UV observations, nor the ?rst to claim evidence of IMF variations at the upper end. Buat, (1987) …Our contribution to the ?eld is to return to the basic method…. Buat, (1987) *

    So if “Star-Birth Myth Shattered” is actually 20 years old!!!

    As for the importance of knowing initial mass function (IMF), Baut (1986) summarises the essence of this story, and the associated paper of 2009;

    The distribution of stellar masses at birth (or IMF) is a basic function of the process of star formation; it is a constraint on theories of star formation and an input parameter for evolution models of stars.

    Pity the article says this so indirectly.

    * REFERENCE COMMENT: Buat, V., Donas, J., & Deharveng, J.M. ;, “The initial mass function for massive stars – A comparison between the total H-alpha and ultraviolet fluxes of a sample of spiral and irregular galaxies“; A&A., 185, 33 (1987) [ See…185…33B . I suggest those interested might read the 1. “Introduction” [ and especially the first two paragraphs.] I also note, that Buat, says in THEIR conclusion;
    It is found that the IMF for massive stars does not vary much from galaxy to galaxy, in agreement with the general impression left by… Scalo (1986)
    Note: So this Buat conclusion is a experimental conclusion by real observations and not a general assumption. They, however, had already assumed that the local IMF did vary depending on the scale in question!

  12. There are characters missing in the second quote given above . It should read;

    “Ours is not the first study to probe the nature of the IMF using H-alpha and UV observations, nor the first to claim evidence of IMF variations at the upper end. Buat, (1987) …Our contribution to the field is to return to the basic method…. Buat, (1987)” *

    (The quote was written in LaTeX, hence the missing characters.)

  13. The link to the origin of this recent paper is interesting. It date to 10th April 2009, but the Media Release by the CSIRO dates to 21 August 2009.
    Also, the information on the galaxies, two taken from images leading this page at the Australian CSIRO site. It has the following statement about these two galaxies in question at; .

    “False-colour images of two galaxies, NGC 1566 (left) and NGC 6902 (right), showing their different proportions of very massive stars. Regions with massive O stars show up as white or pink, while less massive B stars appear in blue. NGC 1566 is much richer in O stars than is NGC 6902.
    The images combine observations of UV emission by NASA’s Galaxy Evolution Explorer spacecraft and H-alpha observations made with the Cerro Tololo Inter-American Observatory (CTIO) telescope in Chile.
    NGC 1566 is 68 million light years away in the southern constellation of Dorado. NGC 6902 is about 33 million light years away in the constellation Sagittarius.”

    There is also a similarly written story as a podcast on the CSIRO site by Dr Baerbel Koribalski, which is very good and might elaborate to some a more useful slant on this story.

    In the end, I have the real suspicion that some of the whole material is basic propaganda for radio astronomy – which is difficult to relate to the general public as it is neither intuitive nor easily digestible.

    An example of the propaganda is at the end of the CSIRO Media Release, stating;

    “The Australian SKA Pathfinder, the next-generation radio telescope now being developed by CSIRO, will find neutral hydrogen gas in half a million galaxies, allowing a comprehensive examination of star-formation in the nearby universe.

    Ummm. Perhaps there is some method to the madness here – but I leave that decision and conclusion to you.

  14. HSBC, may I ask, if you have some expertise in this field?
    I haven’t read so many interesting and informative posts from you for quite a while, which I regret as I see here and now.

    Nereid, thanks for the flowers, but I haven’t read much of the paper (of both papers actually). I merely flew over them. Especially the 16 pages of the paper that is discussed here seemed a bit too much for me, because I didn’t deem it to be so important for the discussion.

    Still, I second HSBC’s comments concerning general aspects of the IMF. It is obviously clear that star formation depends on the region the stars are formed in. That also includes the time. Pop-III stars are thought to have been very massive – more massive than any known stars today. So the IMF must have looked different in previous times.

  15. Thanks to all who posted above (and with some very thought provoking links and comments). IVAN3MAN’s link to the original paper did seem at odds as to what I presumed was common knowledge to astrophysicists wrt the IMFs in various galaxies and didn’t seem to jive with the CSIRO PR, as others have noted. I was also heartened too see no mention from readers thinking this paper may hold clues to the mystery of DM. The universe was created with a fixed amount of matter and energy. In this case no new ‘missing matter’ was implied, only how scientists now view of its’ distribution in the universe. Congrats to Brian and all who have posted (up to now) for not mentioning this non-issue. This combination of great science writing and many insightful comments is what makes UT my first site I visit wrt in-depth stories on astrophysical phenomenon. Great job, keep it up!

    BTW, any one here that can make the launch of STS-128 a reality at 11:59 EDT Friday night 🙂

  16. Ah, maybe it’s the glass of white wine I’m currently enjoying [quite likely, in fact], but I have this terrible suspicion that HSBC’s comments opened up and illuminated the over all dynamics of star formation (and its distribution) in a format I could actually digest. (Modulo all those pesky details and technical papers that I will gladly overlook for now.)

    The scary thing is, of course, that I’m a layman here. 😀

  17. So what’s yours TL (OM)? Mine’s a sauvignon I’m trying for the first time; pretty good! Salud/Prost/Kampai/Ganbei/Bottoms-up/Skoal/… no wait, one does NOT enjoy whites by knocking them back …

  18. Dr. Flimmer said;

    >BLOCKQUOTE>HSBC, may I ask, if you have some expertise in this field?
    I haven’t read so many interesting and informative posts from you for quite a while, which I regret as I see here and now.

    Yes I do. I have some liberal experience regarding stellar evolution, especially associated with stars, binary stars, angular momentum dissipation, and early development of galaxy evolution – much actually being centrally focussed on galactic IMF and SFR . I did published a quite specific paper on galaxy evolution in late 1980’s, but have had little do with it in recent years.
    I usually say very little about my own interests or research work, as few really would comprehend its importance or its implications. With dangerous idiots like Anaconda running about, and other EU fanatics blabbering on endlessly on this EU / PC mumbo-jumbo, who would blame me (or anyone) in not saying more!
    Frankly over the years I’ve grow very tired of being constantly the harmless giving educator – just keeping on feeding the minions – yet gaining little new knowledge for myself. For this reason, I am happy to remain just anonymously distant and well hidden incognito.

  19. Dr. Flimmer said;

    HSBC, may I ask, if you have some expertise in this field?
    I haven’t read so many interesting and informative posts from you for quite a while, which I regret as I see here and now.

    This failure was in the HTML syntax. Sorry

Comments are closed.