One of the first of the second stars


The chemical content of a star that was born relatively shortly after the formation of the Milky Way calls into question conventional understanding of how stars formed in the early Universe.

How did the first stars form early in the history of our Galaxy and the Universe, and what were these stars like? The Big Bang produced only hydrogen, helium and some lithium, so the first stars would have contained only those elements. Given that the stars became extinct long ago, they were probably quite massive — the more massive a star (with masses more than 80 times that of the Sun a possibility), the shorter its lifetime (as little as a few million years). However, we can learn much about the early conditions in our Galaxy, including the types of element formed and the nature of these first stars, by studying surviving stars from a second generation of stars, which formed from the debris of the first. These surviving 'second stars', which are less massive than the Sun and live for many billions of years, can be found in the Galactic halo. Writing in Astronomy & Astrophysics, Caffau et al.1 describe a Galactic halo star that is one of the earliest members of this second generation.

This new work is an extension, and expansion, of Caffau and colleagues' initial report2 on this star, which is catalogued in the Sloan Digital Sky Survey and known by the non-sexy name of SDSS J102915+172927. The authors used the Very Large Telescope (VLT) in Chile to undertake a detailed spectroscopic analysis1 of the elements contained in the star, which is in the centre of the constellation Leo. They report that the star has an extremely low abundance of iron — approximately 1/130,000 that of the Sun. Astronomers refer to a star's iron abundance as its metallicity. This metallicity generally correlates with Galactic time: iron is produced in exploding stars, known as supernovae, and its abundance in our Galaxy has increased progressively over time. Thus, the most iron-deficient — or extremely metal-poor (EMP) — stars are among the oldest, having formed relatively soon after the Galaxy itself, early in the history of the Universe.

And yet SDSS J102915+172927 does not have the lowest known iron abundance. Two other stars3,4 are even more iron-poor, implying that they are older than the star in Leo. Compared to the Sun, however, both stars have a large abundance of carbon and nitrogen relative to iron, as has been noted in several other EMP stars. By contrast, Caffau et al.1 found that carbon and nitrogen were not enhanced in SDSS J102915+172927 — the abundance of these elements with respect to that of iron is consistent with solar values. Similarly to some other EMP stars4, the abundance of oxygen could not be measured in SDSS J102915+172927. However, Caffau and colleagues also defined a metal-mass fraction based on the total abundance of all of the elements heavier than helium, rather than just that of iron. They report that SDSS J102915+172927 has the lowest such value ever measured, and argue that this makes its composition similar to that of the primordial gas that existed shortly after the Big Bang.

The types and abundances of these elements in old, low-metallicity stars are crucial to our understanding of what happened before the stars' formation. First, it is a puzzle how a low-mass star such as SDSS J102915+172927 (which is less massive than the Sun) even formed early in the history of the Galaxy at a time when high-mass stars would seem to be more common. Observational and theoretical studies5,6 have suggested that elements such as carbon or oxygen are necessary for cooling (low-mass) parent gas clouds sufficiently for them to eventually collapse and form low-mass stars. The difference between SDSS J102915+172927, which has a relatively low carbon abundance, and carbon-enhanced metal-poor (CEMP) stars calls into question what is normal for these early stars7. But with so few stars of this type observed, it is difficult to discern a pattern.

Perhaps SDSS J102915+172927 is older than the other EMP stars observed: the relationship between iron abundance and time might not be entirely linear so early on, and thus the lowest metallicity star might not be the oldest. SDSS J102915+172927 might fall into the transition region between the first generation of stars (sometimes referred to as Population III) and the second generation, or Population II; halo, EMP and CEMP stars belong to the latter group.

Or perhaps SDSS J102915+172927 formed in a region of the Galaxy that had particularly low levels of elements heavier than helium. Although carbon can be produced internally in ageing (giant) stars, this star is probably an (unevolved) main-sequence star — a stellar phase similar to the current state of the Sun — and not a giant or a sub-giant2. This means that the carbon and other heavy elements observed in SDSS J102915+172927 must have been synthesized in a supernova and then incorporated into the gas that would form new stars.

The chemical-abundance pattern observed in SDSS J102915+172927 is consistent with what is predicted for such a supernova event2. Yet this star in Leo does not have enhanced carbon, in sharp contrast to one of the unevolved carbon-enhanced EMP stars, which also has detectable strontium4 — a rare heavy element made only in supernovae by a series of neutron captures.

A large scatter in the abundance of heavy, neutron-capture elements with respect to iron is observed for metal-poor stars. (There are many metal-poor halo stars but only a few EMP stars.) This scatter suggests that, at early times, the Galaxy was an unmixed, or non-homogeneous entity, with individual element-synthesis events — that is, supernovae — scattered throughout the halo8. Thus, the greatly varying carbon abundances in these early stars might also result from this heterogeneity. Clearly, observations of additional stars will be needed to probe this early phase of the Milky Way.

Further support for this early history of the Galaxy was provided by Caffau and colleagues' measurement1,2 of the lithium abundance in SDSS J102915+172927. Lithium, produced in the Big Bang, is expected to be at a uniform and primordial abundance level (denoted as the Spite plateau9) in these early stars. Surprisingly, however, the measured lithium abundance in this, and another4, EMP star is low, below the observed Spite plateau. This suggests, at least for some of these early stars, that there is probably a lithium-destruction mechanism occurring during star formation. It will be crucial to find additional evidence of variations in lithium abundance to better understand the earliest stages of star formation.

There will be more stars such as SDSS J102915+172927. Only a small fraction of the thousands of stars in the SDSS database has been observed with the VLT. Nevertheless, this new discovery is a valuable first step in filling in the gaps in our knowledge of the early history of the Universe, and of how stars and elements were formed.


  1. 1

    Caffau, E. et al. Astron. Astrophys. 542, A51 (2012).

    Article  Google Scholar 

  2. 2

    Caffau, E. et al. Nature 477, 67–69 (2011).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Christlieb, N. et al. Nature 419, 904–906 (2002).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Frebel, A. et al. Nature 434, 871–873 (2005).

    ADS  CAS  Article  Google Scholar 

  5. 5

    Bromm, V. & Larson, R. B. Annu. Rev. Astron. Astrophys. 42, 79–118 (2004).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Frebel, A., Johnson, J. L. & Bromm, V. Mon. Not. R. Astron. Soc. 380, L40–L44 (2007).

    ADS  Article  Google Scholar 

  7. 7

    Norris, J. E. Nature 477, 37–39 (2011).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Sneden, C., Cowan, J. J. & Gallino, R. Annu. Rev. Astron. Astrophys. 46, 241–288 (2008).

    ADS  CAS  Article  Google Scholar 

  9. 9

    Spite, M. & Spite, F. Nature 297, 483–485 (1982).

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to John Cowan.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Cowan, J. One of the first of the second stars. Nature 488, 288–289 (2012).

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.