Astron. Astrophys. 346, 798-804 (1999)

1. Introduction

One of the methods called for estimating the age of the Galaxy is based on the analysis of observed abundance of some long-lived radionuclides. The most studied cosmochronometries involve or the actinides , and . Though some promising results have recently been achieved in decreasing the uncertainties affecting the cosmochronometry, the predictions based on the trans-actinides can still be regarded as relatively poor (Arnould & Takahashi 1999). Nevertheless, the recent observation of r-process elements, including Th, in ultra-metal-poor halo stars, such as CS 22892-052 or HD 115444 (Sneden et al. 1996, 1998) has brought some renewed excitement in the estimate of the age of the Galaxy on grounds of the Th cosmochronometry (Cowan et al. 1997; Pfeiffer et al. 1998). With a metallicity as low as [Fe/H] = -3 and a composition enriched in some pure r-elements, these stars provide strong evidence that the production of heavy elements by the r-process already took place early in the history of the Galaxy. Moreover, the abundance pattern of the 15 r-elements heavier than Ba at the surface of CS 22892-052 (or the 9 elements in HD 115444) shows a striking similarity with the solar system r-abundance distribution, leading to the tempting (though hazardous) conclusion that the r-process mechanism is "unique", i.e any astrophysical event producing r-elements gives rise to a solar-like abundance distribution. This conclusion has been critically analyzed by Goriely & Arnould (1997) who showed that this assumption may be valid indeed, but is by far not the only possible one , as the observations in the limited range are equally compatible with an abundance distribution that does not fit the solar one outside the observed domain. This ambiguity is assigned to the fact that the observed CS 22892-052 pattern of abundances reflects primarily nuclear physics properties, and not one or another specificity of a blend of r-process events. This universality assumption is a fundamental prerequisite to build a Th cosmochronometry upon the abundance analysis of metal-poor stars at the present time. In principle, it could be possible to derive the abundance of Th ingested in these metal-poor stars from theoretical extrapolations based on direct fits to the observed abundances. However, in practice, this exercise is affected by uncertainties even greater than when basing the fits on the solar abundances, because of the restricted number of elements observed, the impossibility to distinguish isotopic ratios and the smaller precision in the abundance determination compared with the data available in the solar system. In particular, Goriely & Arnould (1997) showed that the r-elements distribution at the surface of CS 22892-052 could be reproduced satisfactorily by a random superposition of canonical r-process events. In this case, the theoretical extrapolation to the actinide region based on parametric r-process models is simply meaningless. Nevertheless, future accurate observations of r-elements in ultra-metal-poor star could change this situation.

For this reason, we will consider in the present paper the universality assumption to be valid in order to analyze if, despite this difficulty, the recent accurate observation of Th at the surface of ultra-metal-poor stars can indeed provide a reliable estimate of the stellar age by comparing it with the universal r-abundance of Th. Such a procedure requires the estimate of the Th by r-process models, which are known to suffer from very many astrophysics and nuclear physics problems, in spite of much recent theoretical and experimental effort. In this respect, the Th problem is particularly acute, since with U, Th is the only naturally-occuring nuclide beyond , so that the estimate of Th production relies on extrapolation procedures based on fits to the solar (or stellar) r-abundance distribution. In Sect. 2, a brief description of the adopted r-process models is given in relation to the Th cosmochronometry. In Sect. 3, the various uncertainties affecting the Th production are studied and their impact on the estimate of the stellar age is analyzed. In Sect. 4, it is shown by comparing the solar fits to the stellar r-element distribution observed that future observations of Pb, Bi or U could put the Th cosmochronometry on safer grounds.

© European Southern Observatory (ESO) 1999

Online publication: June 17, 1999
helpdesk.link@springer.de