The "impossible" star, along with a pie chart showing its composition; it is made of mostly hydrogen and helium, with trace amounts of heavier elements. Credit: {link url=""}ESO/Digitized Sky Survey 2{/link}

In the beginning, the only elements that existed were hydrogen, helium and very small amounts of lithium. All of the other elements in the period table came later and, rather than forming out of the primordial soup of sub-atomic particles that existed shortly after the big bang, the elements from lithium up to and including iron, were made in the nuclear furnaces at the centres of stars. But all the stars we’ve ever seen have contained some of these heavier elements, a fact which raises the question: were there ever any stars made only of those three first elements?

Astronomers call any element heavier than helium a "metal". That's an awful lot of elements — in fact, 116 of the 118 elements we have discovered so far fall into this category. Astronomer's also talk about the "metallicity" of a star — the fraction of elements in the star, by mass, that are metals. Metallicity is an indicator of a star’s age. Or, more accurately, an indicator of how many previous generations of stars have lived, died and been recycled to make the new star.

Star populations were named in the order the were discovered: I, II and III. Population I stars were discovered first, but created last, and are the youngest stars. Because of this, they have the highest metal content. The Sun is a population I star. Population II stars have lower metal content than population I stars. Population III stars hypothetically have no metal content at all, but none have been observed. Yet.

This poses a problem.

It has been suggested that low-mass stars should not be able to form in the early universe, because the primitive interstellar medium — the gas and dust made of hydrogen, helium and trace amounts of lithium that filled the early universe and still exists between stars today — did not contain enough metals.

To solve both the problem of where the first metals came from and why we haven’t seen any stars without them, astrophysicists have suggested that perhaps only high mass zero-metallicity stars formed. These stars would have had masses around a hundred times that of the sun and their lives would have been over in a (relative) flash, thanks to an inverse relationship between stellar mass and lifetime. When these stars neared the end of the lives, they will have been able to fuse the first 26 elements in the periodic table — up to iron — in their cores. Once they had exploded in supernovae, the newly formed elements would have been spread far and wide throughout the universe, and added to the mix when new stars formed.

These stars would have pushed up the metal content of the interstellar medium and allowed stars with lower masses and higher metallicities to form. This second generation of stars would be population II.

But (isn’t there always a “but”?) a group of astrophysicists, lead by Elisabetta Caffau from Heidleberg University in Germany, recently announced that they have found a low-mass star that also has a low metallicity.

The star resides in the Galactic halo and has a metallicity lower than any other star ever observed. Put simply: this star, which goes by the name “SDSS J102915+172927”, shouldn't exist. Caffau and her colleagues published their findings earlier this month in Nature.

The star gets the first part of its name because it was catalogued in the Sloan Digital Sky Survey (SDSS), and the second part because of where it is in the sky. Simone Zaggia, a co-author on the paper, made a nice animation showing the path of the star in the Milky Way.

Caffau’s team analysed the composition of the star using the spectrographs X-Shooter and UVES of ESO's Very Large Telescope (VLT) in Chile. Spectrographs are instruments that split the light from a star into colour components to find out how much of each element the star contains. When they did this, they found that the star contains no carbon, oxygen or nitrogen.

Low-mass, low-metal stars have been seen before, but these tend to be rich in carbon, nitrogen and oxygen. Carbon and oxygen are thought to be key to the formation of low-mass stars, because they cool the gas and dust in the interstellar medium during star formation. The existence of a low mass star without carbon or oxygen seems to indicate that what we think are the necessary levels of these two elements for low-mass star formation are, in fact, not necessary at all. If this were the case, it would mean a dramatic increase in the diversity of stars in the early universe.

Caffau and her team don’t think that SDSS J102915+172927 is an anomaly. They expect that between 5 and 50 stars from those that can be analysed by the VLT, and even more in the whole SDSS sample, will have similar or lower metallicities to the new most primitive star.

This newly discovered star may indicate that, if we dig a little deeper, we could discover that, in terms of diversity, the oldest stars in the universe are not so different to the youngest ones after all.


Caffau E, Bonifacio P, Franois P, Sbordone L, Monaco L, Spite M, Spite F, Ludwig HG, Cayrel R, Zaggia S, Hammer F, Randich S, Molaro P, & Hill V (2011). An extremely primitive star in the Galactic halo. Nature, 477 (7362), 67-9 PMID: 21886158