≡ Menu

Red Dwarfs: A Rich Harvest

I never have trouble finding topics to discuss on Centauri Dreams, but this morning’s take was unusually bountiful. For the past several days I’ve had two embargoed stories to choose from, both going public this PM. Do I write about tripling the number of stars in the universe, or do I choose the first analysis of a ‘super-Earth’ atmosphere? It’s a tough choice, but I’m going with the stars, given that the story relates to what I consider the most fascinating venue for astrobiology, planets around red dwarfs. We’ll do the super-Earth atmosphere — fascinating in its own right — tomorrow.

The story comes out of Yale University, whose Pieter van Dokkum led the research using telescopes at the Keck Observatory in Hawaii. We’ve long known that because of their faintness and small size, getting a handle on the red dwarf population was problematic. Usually, I’ve seen a figure around 75 percent cited for the Milky Way, meaning most stars in our galaxy are red dwarfs (the Sun, a G-class object, turns out to be representative of only about seven percent of main sequence stars, meaning we live around a star that is not typical). And because red dwarfs are so abundant, the consequences for astrobiology are obvious if we determine habitable planets can orbit them.

The Yale team has looked at red dwarfs not in our galaxy but in eight relatively nearby elliptical galaxies, located between 50 million and 300 million light years away. What they discovered is that there are about twenty times more red dwarfs in these elliptical galaxies than in the Milky Way. Says van Dokkum:

“No one knew how many of these stars there were. Different theoretical models predicted a wide range of possibilities, so this answers a longstanding question about just how abundant these stars are.”

Elliptical galaxies make up between ten and fifteen percent of the galaxies in the local universe, and the finding triples our best guess about the total number of stars in the universe, thereby increasing the number of planets we assume to be orbiting these stars. We’ve seen robust planetary systems around stars like Gliese 581 and can assume similar systems exist in the galaxies under observation. The red dwarfs recently discovered are typically more than ten billion years old, giving life plenty of time to gain a foothold. Indeed, van Dokkum talks about ‘possibly trillions of Earths orbiting these stars,’ a notion that gives still more punch to the Fermi paradox.

Image: Galaxies in the cluster Abell S0740, over 450 million light-years away in the direction of the constellation Centaurus. The giant elliptical ESO 325-G004 looms large at the cluster’s center, as massive as 100 billion of our suns. Hubble resolves thousands of globular star clusters orbiting ESO 325-G004. Globular clusters are compact groups of hundreds of thousands of stars that are gravitationally bound together. At the galaxy’s distance they appear as pinpoints of light contained within the diffuse halo. Other fuzzy elliptical galaxies dot the image. Some have evidence of a disk or ring structure that gives them a bow-tie shape. Several spiral galaxies are also present. The starlight in these galaxies is mainly contained in a disk and follows along spiral arms. Credit: NASA, ESA, and The Hubble Heritage Team (STScI/AURA).

If a multi-billion year old civilization existed in a typical galaxy, would it have made it all the way to Kardashev Type III status, able to put to use the entire energy resources of its galaxy? If so, wouldn’t we see some sign of its handiwork? Dick Carrigan has been studying such issues for years, looking for Dyson spheres by sifting through Infrared Astronomy Satellite (IRAS) data for objects that radiated in the infrared and carried the signature of a Dyson sphere, in which a star is surrounded by a swarm of energy-catching habitats or even completely enclosed within their shell.

No luck yet, but what Carrigan calls ‘interstellar archaeology’ gets more and more interesting when we consider a tripling of the number of potential life-giving stars in the universe (see Carrigan’s analysis of Kardashev Type II and III signatures and his Dyson sphere methodology here). What kinds of traces would a Type III civilization leave, and would we recognize it if we saw it? We do know that elliptical galaxies are made up of older, low-mass stars and show little star formation activity compared to more active spiral galaxies. Perhaps a lack of discernible Type III activity is telling us that technological civilizations have a relatively short lifetime.

Beyond these blue-sky musings, though, a tripling of the stars in our universe could have an impact on our understanding of how galaxies evolve, forcing us to take account of the mass of a much larger population of red dwarfs than we thought existed, and thus providing a new constraint on dark matter and its effects in relation to elliptical galaxies. That’s a helpful outcome, and it will be fascinating to see how the results of this paper are received and put to use.

The paper is van Dokkum and Conroy, “A substantial population of low-mass stars in luminous elliptical galaxies,” published online in Nature 1 December, 2010 (abstract)


Comments on this entry are closed.

  • James M Essig December 1, 2010, 15:01

    Hi Paul;

    Such a galaxy could contain as many as 6 trillion to 12 trillion red dwarfs.

    It seems that nature tends to produce about 75 percent of all stars as red dwarfs.

    The number of red dwarfs in the visible universe could be about 10 EXP 24. Assuming that one in 10 of these stars has an Earth like or terraformable planet, the number of persons such planets could support including future human explorers, indigenous ET persons, and any bodily UT persons is enourmous.

    Say that each planet could support 10 billion persons or perhaps as many as 100 billion by completely green and sustainable infrastructure, a total of 10 EXP 35 persons may be supported simultaneously. Assuming a life expectancy of 1,000 years, over the lifetime of low mass range red dwarfs or about 10 trillion years, 10 EXP 45 persons could be born and live fulfilling lives.

    If we learn superluminal travel methods, the number of persons that can have the opportuinity to live is (10 EXP 45)[(nC/C)] where n is a greater than one multiplier of the speed of light. This of course neglects space time expansion, however as the limit of superluminal travel somehow approaches infinity, than for any finite level of space time expansion and finite increase in its rate of expansion, the expansion of the universe would be a trivial consideration for finite travel distances no matter how large.

    I do not expect infinite C travel anytime soon, however, I think we will see gamma factors in the 2 to 5 range within the next few centuries or speeds of 0.866 C to 0.98 C.

  • Adam December 1, 2010, 16:24

    Hi Paul

    Perhaps K III Civilizations are smart enough to figure out how to fill and control their galaxies without damming the light of all the stars? Alternatively perhaps they fill the spaces between the stars because that’s where all the free-floating mass is? In otherwords, too many options to know… a chronic issue for SETI & SETA.

  • Parker Shaw December 1, 2010, 16:40

    So, about this hot topic of dark matter/dark energy stuff, is that possible that there is just enough NORMAL matter that we haven’t detected?

  • Paul Gilster December 1, 2010, 16:50

    Parker, it seems unlikely. If there were enough normal matter in play that we haven’t yet detected, it should be all over the place, and various attempts have been made to find it. Evalyn Gates gets into this issue in her book Einstein’s Telescope — I don’t have my notes where I am now, but the idea is that missing normal matter should turn up if, for example, you study the Magellanics and look for occultations by dark objects between us and the satellite galaxies. Statistically, that sets up the ability to constrain the amount of dark matter such objects could represent, and it doesn’t add up to nearly enough to explain gravitational lensing in galactic clusters or, for that matter, the problem of galactic rotation. When I get back to my desk, I’ll see if I can find some references for you.

  • Ronald December 1, 2010, 17:52

    If I understood well, elliptical galaxies are generally old and metal-poor. This could mean that the abundance of (terrestrial) planets in them may be disappointing.

  • Mike December 1, 2010, 18:14

    “typically more then 10 billion years old,”
    These M-Dwarfs are likely to be very metal poor if they’re that old. That also means a metal poor proto-planetary disk. If terrestial planets are still able to form will there be enough heavier elements present for life to arise?
    These Population 2 stars are likely deficient in all elements heavier then helium . Ofcourse there aught to be huge numbers of younger M-Dwarf stars present in these ellipticals as well. However, every M-Dwarf that has ever formed should still be in existence. They last a long time.
    The older ones will out number the younger. Roughly speaking, the metal poor older M-Dwarfs will out number the metal rich younger ones. What all this means is while we may have discovered that there are vastly more M-Dwarf stars in the universe this may not necessarily mean that most of these star’s could be abodes for life.
    What ever the debate about how suitable M-Dwarf stars in general could be as a parent star to a habitable planet, if there is not enough Silicon or Iron or Nitrogen present to form rocky planets or to provide the chemical basis for even simple life because these ancient M-Dwarfs are metal poor then I think that is a real show stopper.
    Perhaps future studies of these newly discovered stars may give us a better idea of their metallacity.

  • andy December 1, 2010, 19:33

    @Ronald: depends which elliptical galaxies you are talking about. Galaxies like M87 which lies at the heart of the Virgo cluster do contain stellar populations of substantial metallicity.

    @Mike: As for old stars being metal-poor, I refer you to this paper about stars in NGC 5128 (Centaurus A):

    Two burst models, with 70-80% of the stars formed 12+/-1 Gyr ago and with 20-30% younger contribution with 2-4 Gyr old stars provide the best agreement with the data. The old component spans the whole metallicity range of the models (Z=0.0001-0.04), while for the young component the best fitting models indicate higher minimum metallicity (~1/10 – 1/4 Z⊙). The bulk of the halo stars in NGC5128 must have formed at redshift z≳2 and the chemical enrichment was very fast, reaching solar or even twice-solar metallicity already for the ~11-12 Gyr old population.

    At least some galaxies appear to have managed to form high-metallicity stars very early on…

  • Paul Gilster December 1, 2010, 21:02

    Re Parker Shaw’s question on whether normal matter could account for ‘dark’ matter:

    Parker, I’ve gotten back to the office, where I’m looking at Gates’ book. Bohdan Paczynski (Princeton) was interested in the idea of MACHOs (massive compact halo objects) as candidates for dark matter — these might be anything from very dim stars to black holes and even Jupiter-class planets. Going on the assumption that MACHOs would, even if not visible to us, warp spacetime because of their mass, he reasoned that the deflection of light in this gravitational lensing would be visible to us. In his proposed experiment, he used the Large Magellanic Cloud, knowing that because MACHO masses would be much less than the Sun’s, their lensing would produce only a subtle signal.A white dwarf in front of a star that is 170,000 light years away in the Large Magellanic Cloud will lens the light of the star and produce two images of it, with a tiny separation (making this ‘microlensing’ — we wouldn’t see two images because of the tiny angle of separation, but we would be able to detect a brighter image of what appears to be a single star).

    Thus the idea came down to searching the LMC for stars that are brighter than they would have been if they weren’t observed through the MACHO lens. A MACHO moving through the halo must occasionally pass in front of a star in the LMC, making it appear brighter, then fading back to original brightness as the MACHO goes past it. Three teams of scientists went to work on this problem in the 1990s (Gates outlines all this in her book). The teams did find MACHO candidates, and by the end of the decade had a combined total of 20 microlensing events from observations in both the Large and Small Magellanic Clouds.

    And here’s the crucial part, from the Gates book: “The final paper of the MACHO collaboration, published in 2000, concluded that a galactic halo consisting entirely of MACHOs was not ruled out, and estimated that about 20% of the galactic halo was in the form of MACHOs. The EROS team preferred to present its results as an upper limit on the number of MACHOs in the halo, with no more than about 8% of the halo in MACHOs having masses of about one-tenth to one times the mass of the Sun. A combined analysis of the two experiments showed that, within the uncertainties of each experiment, they are consistent with each other and that less than 20% of the halo is in the form of MACHOs… MACHOs, the least exotic candidates for dark matter, have now been effectively ruled out as the main component of the dark matter, leaving WIMPs to dominate the galaxy.”

    WIMPs are weakly interacting massive particles, the other major model for dark matter. Anyway, as you can see, I was wrong in thinking that it was merely occultation that these teams were looking for — they were counting on actual lensing to help them make their calculations. For a detailed look at all this, check the Gates book out, and I’ll also be reviewing a new book that gets into this in coming weeks.

  • kurt9 December 2, 2010, 2:20

    So, there are more red dwarfs than expected. NASA’s supposed to announce the discovery of some bacteria that lives off of arsenic that someone dredged up from Lake Mono in California. No doubt that life will be primitive (pro-karyote). I still stand by my prediction:



    Eukaryote life is still rarer than hens’ teeth and so are Oxygen atmosphere planets.

  • Ronald December 2, 2010, 8:56

    As I read in Brian Wang’s Next Big Future (my other favorite website), the total number of stars in the universe is now guesstimated at some 3 * 10^23, see http://nextbigfuture.com/2010/12/discovery-triples-number-of-stars-in.html#more.

    As one commenter notes that is about half a mol (Avogadro’s number, just over 6 * 10^23). Double it again (see my next observation) and there could be a mol of stars in the observable universe.
    Incidentally, the *entire* universe is estimated, by Alan Guth’s Inflation Theory, to be at least 10^23 times as large as the observable universe.

    Hm, maybe there is indeed some special significance to the number 23 (the 23 enigma) ;-)

    Observation: maybe spiral galaxies, such as our own MW, also contain more red dwarfs than previously suspected in their halo’s. Individual red dwarfs (i.e. not in clusters) in those outer regions could be hard to detect.

  • Ronald December 2, 2010, 9:05

    @kurt9: interesting, but as we discussed in a recent thread, oxygen rich atmospheres could also be formed with only prokaryotes. The Great Oxygenation Event on earth did probably not (primarily) take place because of the rise of Eukaryotes, but because of saturation of oxygen sinks in the ocean and crust.
    The positive side of this, if indeed correct, is that we could then detect any planet with (O2 based) photosynthetic life by means of spectroanalysis of its atmospheric O2/O3 content, the downside being that this way we will then not be able to distinguish the planets with higher life.

  • Mike December 2, 2010, 12:59

    Thank you Andy for the link to that interesting paper. It looks like it is well worth the read.

  • Parker Shaw December 2, 2010, 13:57

    Paul, thanks a lot for the information!

  • amphiox December 2, 2010, 14:10

    @Ronald: Indeed, it is far more likely that The Great Oxygenation Event enabled the rise of the eukaryotes, and not the other way around.

  • kzb December 2, 2010, 14:50

    I am puzzled as to why the IMF for these elliptical galaxies is said to be different to that of the Milky Way, based on these figures.

    For example, if you look at the RECONS site (link below -and incidently this has been recently updated after I thought it had gone dead), you will see that within 10pc, class M stars make up 72% of the stars class M and above.

    Then bear in mind this RECONS census is almost certainly incomplete (and the imcompleteness is in class M and lower), there would not seem to be a significant difference from the MW stellar make up in this regard.


    The other point I’d like to make is about metallicity and terrestrial planet frequency. There is NO EVIDENCE that lower metallicity equals lower terrestrial abundance. In fact many formation models give precisely the opposite result !

  • andy December 2, 2010, 15:14

    Eukaryote life is still rarer than hens’ teeth and so are Oxygen atmosphere planets.

    Well of course eukaryotic life is vanishingly rare in the universe, as any alien life would by definition NOT be eukaryotes. :-)

    As for oxygen-atmosphere planets, these may be quite common: seems to be a natural result of the bombardment of ice moons by gas giant radiation belts. Oxygen atmospheres are also predicted on ocean planets.

  • Adam December 2, 2010, 16:16

    Years ago Jim Kasting pointed out at least two ways oxygen atmospheres could accumulate without life’s involvement – cold, icy planets big enough to retain it, and hot wet ones undergoing ocean-loss via photolysis. Seems the former is more common than the latter, but we’ll need to watch out for astrobiological false positives from both. I can’t but help think that Fermi’s Paradox has something to do with oxygen’s arising and the timing of its origin. A bunch of old red dwarfs with a bevy of possible planets, yet no ETAs in evidence, just underlines that for me. I suspect there will be lots of deep-freeze Earths around such stars, but very few with life powered by unhindered starlight.

  • kurt9 December 3, 2010, 13:54

    but as we discussed in a recent thread, oxygen rich atmospheres could also be formed with only prokaryotes. The Great Oxygenation Event on earth did probably not (primarily) take place because of the rise of Eukaryotes, but because of saturation of oxygen sinks in the ocean and crust.

    That’s not what one of the papers I linked to suggests. The graph suggests otherwise.

    Well of course eukaryotic life is vanishingly rare in the universe, as any alien life would by definition NOT be eukaryotes.

    Nick Lane makes a convincing “design space” argument for why the Eukaryote is necessary for all advanced life.

    Oxygen atmospheres may be common, but they would require an active mechanism to maintain them. Oxygen is very reactive and will react with almost anything. Moons orbiting gas giants may have Oxygen atmospheres as result of the ionizing radiation, but we have no examples of these in our own solar system. Water worlds (we have already found one of these) will have some Oxygen resulting from the breakdown of the water vapor from the solar radiation. Its not clear to me just how much Oxygen can be produced this way.

    If you are right, there will be a lot more habitable planets than there would be if I’m right.

  • andy December 3, 2010, 18:30

    Nick Lane makes a convincing “design space” argument for why the Eukaryote is necessary for all advanced life.

    Actually he does nothing of the sort. He suggests that some method that can provide energy for the cell with low cost in terms of number of genes necessary to maintain that structure is necessary for complex life. That is not the same thing as requiring them to be members of the Eukaryota, which are a group of organisms that evolved on this planet. Eukaryotes provide one realisation of this strategy for supplying the organism with energy. Even if an alien biosphere evolved some analogue of this technique to provide energy, they would not be eukaryotes as they would be representatives of a separate evolutionary process distinct from the one that happened on Earth.

  • Ronald December 5, 2010, 12:17

    @andy: of course you are right about the Eukaryote cell type per se being unique to earth, but kurt9 probably means that a similar type and level of energy provision will be necessary. And according to convergent evolution (similar environmental circumstances produce similar forms and functions) something similar to the Eukaryote in functioning may be expected.

    @kurt9: I agree with what you say about non-biogenic oxygen atmospheres: since oxygen is so reactive. although I have no figures ready, I strongly suspect that under more or less earthlike temperature conditions atmospheric equilibrium oxygen levels as a result of non-biogenic processes, particularly photolysis of H2O by UV, will be quite low. Furthermore, such photolysis would result in loss of H2 into space and high CO2 levels.

  • forrest noble December 5, 2010, 22:51

    Hey Jim,

    That certainly is a large quantity of potential habitats. They also might be a great place to find more advanced civilizations since these stars could be much older than the sun. Also the closest star to us, Bernard’s star, falls into this category which would mean that if it had potentially habitable planets it might be a potential human habitat since it orbits a long way out from its companion stars which seemingly could allow close planets to be relatively unaffected by the ternary (three star) system.

    Also the life of a dwarf star would be generally much longer than the sun’s life providing any companion stars would not be involved in the future of the life on its planets.

  • kzb December 6, 2010, 14:56

    There is an arxiv version of the full paper available without pay wall here:

    I’ve only skim -read it so far. One issue is that a universal IMF has been pretty well assumed in many areas. In spiral galaxies for example the tightness of the Tully-Fisher relationship supports this. OK that does not apply to ellipticals but you see where I’m coming from.

    I still say I can’t see what is the massive difference in IMF anyway. 72% around here, 80% there, not so great anyway?

  • Ronald December 7, 2010, 5:55

    @kzb: the cited publication mentions “making up >80% of the total number of stars and contributing >60% of the total stellar mass”.

    So low mass red dwarfs (< 0.3 Msolar) make up *more than* 80% of total number and *more than* 60% of total mass.

    I agree that there still is something remarkable about this conclusion: since the average M dwarf is so much smaller in mass than the sun (in our galactic neighborhood about 0.24 or 0.25 Msol, according to RECONS data) and this research does not even consider all M dwarfs but only the smallest (60%.

    For comparison, our own galactic neighborhood, using RECONS data;
    – M dwarfs make up 72-75% (also depending on criteria for M dwarf) of total stellar number, but only just over 45% of total stellar mass (since the average M dwarf = 0.25 Msol and the average star = 0.4 Msol).
    – Taking only the low mass (< 0.3 Msol) dwarfs, the number abundance is reduced to just over 50%, but the mass fraction to just over 20% of total (their average mass being only about 0.16 Msol).

    In other words, in order for low mass (< 0.3 Msol) M dwarfs to constitute more than 60% of total stellar mass, you really need to have *a lot* of them, well over 85% of all stars in the galaxy. However, if indeed these M dwarfs are some 20 times more abundant in those elliptical galaxies than in our MW, then the numeric and mass % may be even higher.

  • Ronald December 7, 2010, 6:06

    I have a question related to this research: would it be possible to establish the stellar population distribution (mass, spectral type) of an entire galaxy, such as Andromeda, by analyzing the total spectrum of that galaxy? Such as fraction of M dwarfs, K stars, G stars, etc.

  • kzb December 7, 2010, 14:22

    Ronald, thanks for all the figures, not sure if I follow it yet ! You are saying the number average for M dwarfs is not too different, >80% for ellipticals versus 72-75% for the MW, yet the proportion of stellar mass as M-dwarfs is >60% ellipticals and only 45% for the MW?

    How can this be? That would mean the average M-dwarf in the elliptical would be more massive than the average M dwarf in the MW? Isn’t this the reverse of what they are saying?

    Let us also remember, the RECONS sample is probably still incomplete for this class of star. The within-10pc count increased between Jan 2009 and Jan 2010. They are only accepting proven-beyond-reasonable-doubt candidates.

    Something else about RECONS, it depends on high proper motion detection, therefore nearby stars that have low apparent motion, or motion predominantly along the line of sight, won’t be detected (as I understand things, someone on here might know different).

    Therefore I think it is inevitable there are still more nearby M dwarves yet to discover, and that could push up the number proportion to 80%+ here.

  • kzb December 7, 2010, 14:31

    Ronald, your question about stellar mass determination of whole galaxies might be related to this:


    Although precisely how these claims have been calculated I’m not sure, I could not find it.

  • Ronald December 7, 2010, 19:58

    @kzb: my basic assumptions (based on RECONS and a few other surveys) were:
    – The MW galaxy stellar population consists of 75% M dwarfs, the average M dwarf being 0.24 solar mass and the average of all stars being 0.4 solar mass.
    – This results in the M dwarfs of our MW making up about 45% of stellar mass.

    – 85% M dwarfs would then make up about 60% of stellar mass, also because the greater proportion of M dwarfs would bring down average and total stellar mass.

    – If considering only the smallest M dwarfs of < 0.3 solar mass as in the cited study, assuming an average of 0.16 solar mass, the 60% of stellar mass would then be reached at about 87 or 88% of just these *smallest M dwarfs*. In fact this seems like an unlikely population (mass) distribution, strongly skewed towards the smallest stars. And this implies an even much higher proportion of all M dwarfs combined, probably well over 95%.

  • Ronald December 7, 2010, 20:03

    kzb, thanks for the link, but this article just mentions total star number and total stellar mass of Andromeda.
    What I meant is whether a *population distribution* (i.e. relative abundance of different mass classes and/or spectral types) could be derived from the spectrum of the entire galaxy.

  • kzb December 8, 2010, 8:45

    Ronald: OK I follow your reasoning better now. Star formation is meant to be pretty well extinct in ellipticals, so perhaps this simply represents an aged stellar population, rather than the Initial Mass Function?

    I still say, I don’t see why the population is seen as a massive departure from that of the MW. As I said, the RECONS data is probably still incomplete. Just last night I was looking at a recent (2010) article where they had found that a higher than expected proportion of nearby M dwarfs are in fact binaries. So the 75% figure is probably an underestimate already.

    As to your second post on the population distribution in whole galaxies: as I understand things, the mass-to-light ratio is taken as universal. Certainly in the Tully-Fisher relation this needs to be the case or it wouldn’t work. Smaller galaxies plot off the line, but when corrected for extra dust extinction they are back on it.

    Also let us remember, non-baryonic dark matter was invented because the rotation curve of spiral galaxies did not plot with the luminosity decrease as you move out from the galactic centre. So the assumption of a universal stellar mass-to light-ratio is implicated here. If you are allowed to dump the universal M/L ratio, it is quite possible to model galactic rotation without non-baryonic matter.

  • Ronald December 8, 2010, 17:17

    @kzb: since elliptical galaxies are generally old, poor in interstellar matter and hence poor in star formation, then perhaps the high proportion of low-mass red dwarfs is (partly) a result of the larger mass/hotter stars going extinct, leaving the long-lived low-mass red en orange-red dwarfs.
    Ultimately this is also the fate of spiral galaxies, be it much later.

  • kzb December 9, 2010, 14:19

    Ronald: that’s what I said ! All the more reason to question why this IMF is said to be different to that of the MW.

    The paper about nearby M-dwarf binaries is here: