RSS Feeds

Email a Friend Printer Friendly

Measurement of stellar age from uranium decay

February 05, 2001

For the first time, an international team (led by Roger Cayrel, from Paris Observatory), could measure one uranium line in absorption in a star. This observation has several important implications. It is a great discovery, obtained thanks to the high resolution spectrograph UVES, assembled on one of the 8m-diameter telescopes of the Very Large Telecope (VLT) of the European Southern Observatory (ESO) in Paranal (Chile).

Roger Cayrel and his collaborators (among them P. Francois, V. Hill, F. Spite and M. Spite, of Paris Observatory) observe a sample of stars which are the most deficient in heavy elements, produced by the stellar nucleosynthesis (mainly ejecta from supernovae of type II). Their goal is to study the first stars to have been formed in the Universe, because formed from gas not yet polluted by other stars, and then go back to the formation of the Galaxy. The detailed analysis of element abundances in these very primitive objects (certainly more primitive than the clouds of the Lyman-alpha forest observed at the redshift of 4) must allow to better understand the origin of the significant variations of chemical composition from object to object, observed at these very weak metallicities.

The first observations of this program have discovered a giant star CS 31082-001 of metallicity equal to 1/800 that of the sun. On the other hand, deficiency in elements produced by neutron capture (of which uranium), is only of a factor 10 compared to the sun. It is because of these particular abundances that, for the first time, an uranium line is measurable in a stellar spectrum. Even in the sun, it was never possible to detect uranium. Indeed, the nearby line of iron (cf. figure) is then so broad in stars like the sun, that it overflows on (and hides) the line of uranium. The latter, at awavelength of 385.957 nm, is visible on figure 1, which identifies also the neighbouring lines.

The interest of the observation is that the uranium ( 238 U), since its formation, decline with a half-lifetime of 4.47 billion year. If one knows its production ratio with another stable element produced simultaneously by the same process, it is then possible to deduce the time since the production of uranium (undoubtedly that of the supernova explosion that gave rise to all elements beyond lithium present in CS 31082-001), plus the time of light arrival from the star, which is negligible.

Until now only thorium ( 232 Th), has been used to this end, but its half-life time is 14.0 billion year. Its decline is thus of only a factor two at most, and the uncertainty on the age determination, taking into account all other uncertainties, is about 4 to 5 billion year, that is to say a third of the age of the Universe. With uranium, which declines by a factor 8 in the same time, one hopes to have three times less error.

However, before being able to announce an age with this precision, it will be necessary to reduce the uncertainty which still exists, in atomic physics, on the one hand on the oscillator strength of the line (a team from the CEA is mobilized on this determination) and on the other hand on the production ratio between uranium with respect to thorium and the stable elements the most nearby in mass number (from Os to Bi). For the moment the preliminary estimate of the age of the uranium formation, practically the age of the Galaxy (which had then formed only less than one thousandth of the current number of stars), is of 12.5± 3 Gyr. The error should be soon significantly reduced.

The age of old stars is classically estimated from the computation of the time needed for exhausting (by nuclear fusion reactions) the matter in the stellar core, but these computations are affected by some uncertainties in the rate of these fusion reactions and by uncertainties in the internal stellar structure. The independent evaluation of the age of old stars by observation of a stellar atmosphere (close to the age of the Universe) is therefore welcomed.




Observatoire de Paris and CNRS