News & Views | Published:


A hint of normality at last?

Nature volume 477, pages 3739 (01 September 2011) | Download Citation


The chemical diversity of the oldest stars is greater than we thought. The discovery of an extremely iron-poor star with a 'normal' ratio of carbon to iron challenges our perception of early chemical enrichment. See Letter p.67

The chemical abundances of the oldest stars in the Galaxy hold a key to what the Universe's conditions were like at the earliest times. Indeed, the three most chemically primitive stars currently known have abundance patterns extremely different from those formed after the Universe's first billion years — for example, their abundance ratio of carbon to iron is about 10–1,000 times larger than is found at later times, indicating quite different conditions between the two epochs. On page 67 of this issue, Caffau et al.1 report a chemically primitive star that is in some respects more 'normal' than these objects, indicating a larger chemical diversity, which challenges our understanding of the first stars.

Most astrophysicists agree that the Big Bang hypothesis provides the best description of the formation of the Universe. According to the standard version of this theory, a few minutes after the Universe began the only chemical elements were hydrogen, helium and lithium. At that time, some 13.7 billion years ago, their fractions by mass were 0.75, 0.25 and 2.8 × 10−9, respectively2,3. Astronomers, on the other hand, have observed no stars devoid of elements more massive than lithium. The two most iron-poor stars, which are believed to have ages of approximately 13 billion years, have an observed iron abundance about 10−5.5 that of the Sun — a small but well-determined amount. Further, the observed fractional lithium abundance in most stars that formed close in time to the beginning of the Universe, and in which the observed abundances should not have changed from their initial values, is 8.3 × 10−10, some three times smaller than predicted4. Finally, in the three most iron-poor stars (all with less than 10−4.5 solar iron abundance), carbon, nitrogen and oxygen are present in prodigious amounts relative to iron; and in one of them, the abundances of sodium, magnesium and aluminium relative to iron are at least 100 times those of the Sun.

What do these observed abundances tell us? One suggestion is that supernovae (the final explosions of massive stars, which produce essentially all of the elements heavier than lithium and enrich the gas clouds from which later stars form) were very different within the first few hundred million years from those that followed, leading to the very different abundance patterns. Another possibility is that large overabundances of carbon and/or oxygen may have had a crucial role in determining the nature of the first stars to form.

A hint of normality has been restored to the field by Caffau and colleagues' discovery1 of the extremely iron-poor dwarf star, SDSS J102915+172927, which has 10−5.0 the iron abundance of the Sun. This value places it in the same range as the three most iron-poor stars mentioned above. Normality prevails in the sense that this star is not strongly carbon enhanced: the authors did not detect carbon in it, and the carbon-to-iron abundance ratio upper limit is not too different from the solar ratio. As Caffau et al. point out, the low carbon abundance in this object seems inconsistent with the prediction of Frebel et al.5 that large carbon and/or oxygen abundances is an essential ingredient that provides cooling of the gas clouds from which the early low-mass, long-lived stars we observe today were formed.

The question that begs to be answered is: what does 'normal' mean during the Universe's first few hundred million years? Given that the majority of these four most iron-poor stars is carbon-rich, should the carbon-rich stars not be considered normal and the 'carbon-normal' object abnormal? More to the point, should one think in terms of two different types of chemical-enrichment sources that produce the different observed chemical signatures — perhaps the 'mixing and fallback' type of supernova6 for the carbon-rich stars and the standard 'core-collapse' supernovae for the carbon-normal — or does one need something quite different? Another intriguing question is whether the 'carbon-normal' SDSS J102915+172927 is more primitive than the three carbon-rich members of the most iron-poor stars discussed here.

There will be considerable interest in the lithium abundance of this star, which is anything but normal: Caffau et al.1 were unable to detect lithium in the spectrum of SDSS J102915+172927 (despite the expectation that it should be readily detectable) and report a lithium mass fraction of less than 6.8 × 10−11 — more than 40 times smaller than the predictions of Big Bang nucleosynthesis, the process by which atomic nuclei were formed in the early Universe. Only one of the previously known three most iron-poor stars (HE 1327-2326) has an effective temperature at which stellar evolutionary effects are not expected to have greatly altered its original lithium abundance, and this star is also lithium deficient7, by a factor greater than 100. Thus, as far as we know, all four could have been born lithium-poor. Although lithium-deficient stars are not unknown in the Galaxy's halo, which contains the oldest and more iron-poor stars in the system and in which the lithium deficiency is driven perhaps by their being members of binary systems8, they comprise only about 5% of the Galaxy's halo stars. The absence of lithium in the most iron-poor stars discussed here is an exciting and potentially fundamental result. What has happened to the lithium created at the birth of the Universe?

The caveat to the above discussion is, of course, the small number of currently known iron-poor stars that have less than 10−4.5 the solar iron abundance. Caffau et al.1 comment that they expect 5–50 stars of similar (or lower) iron content to that of SDSS J102915+172927 to be found in the Sloan Digital Sky Survey, in which they discovered this star. If they, and other currently planned surveys for the most metal-poor stars, are successful, the long-standing tyranny of small numbers will indeed have been overcome.


  1. 1.

    et al. Nature 477, 67–69 (2011).

  2. 2.

    et al. Astrophys. J. Suppl. 170, 377–408 (2007).

  3. 3.

    , & J. Cosmol. Astro-Part. Phys. 11, 012 (2008).

  4. 4.

    et al. Astron. Astrophys. 522, A26 (2010).

  5. 5.

    , & Mon. Not. R. Astron. Soc. 380, L40–L44 (2007).

  6. 6.

    & Nature 422, 871–873 (2003).

  7. 7.

    , , , & Astrophys. J. 684, 588–602 (2008).

  8. 8.

    , , , & Astrophys. J. 571, 501–511 (2002).

Download references

Author information


  1. John E. Norris is in the Research School of Astronomy and Astrophysics, The Australian National University, Canberra, ACT 0200, Australia.

    • John E. Norris


  1. Search for John E. Norris in:

Corresponding author

Correspondence to John E. Norris.

About this article

Publication history



Further reading


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.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing