Melanges.en

A fundamental tool to explore the formation history of our Galaxy is to observe its oldest stars. Since the interstellar gas from which stars are formed is more and more enriched in heavy elements ("metals") synthetised in the stellar interiors by nuclear reactions, the mean to detect the oldest stars is to observe those who are the poorest in metals. This is the ambitious program undertaken by an international team of researchers, and among them astronomers from Paris Observatory.
In giant low mass stars, convection does not predict any mixing between the atmosphere of the stars and the deep layers where carbon is transformed into nitrogen. However, it seems that such a mixing takes place in some evolved very metal-poor giants...

Generally speaking, convection theory does not predict that the surface composition of a low-mass star should change during its lifetime. Its composition is thus a good tracer of Galactic matter at the time of the formation of the star. In the deep layers nuclear reactions will provide energy to the star, but the products of these reactions will remain in the stellar interior and will not contaminate the surface layers.

An international team including several members of the Paris Observatory (and, in particular the PI) has studied on this basis with the VLT (Large Programme "First Stars" ID 165N-0276) a sample of very old, extremely metal-poor stars (they were born about 13 billion years ago) to constrain the early phases of Galactic evolution.

Most elements in the atmosphere of the giant stars, show constant abundance ratios with a very small scatter, but carbon and nitrogen (two elements very abundant in the Universe) are an exception. For these elements the scatter from star to star is very large. It can reach a factor of 100 for the N/Fe ratio.
Either this dispersion represents a true dispersion of C and N in the interstellar medium from place to place in the early Galaxy (primordial scenario), or the original C and N abundances in the atmosphere of the giant stars have been altered by variable degrees of mixing with the H burning layer (in situ scenario) where carbon is transformed into nitrogen at a temperature of about 2 10⁷K.

How to discriminate between these two scenarios ?

In the deep layers of the star where hydrogen is burnt, carbon is transformed into nitrogen and the abundance of ¹³C relative to ¹²C increases. If, in the sample, there are some stars where a mixing has taken place between the atmosphere of the star and these deep layers we expect in the atmosphere of these stars:
- a strong abundance of nitrogen correlated with a low abundance of carbon, and a low ratio ¹²C/¹³C
- the total abundance C+N to be the same in mixed and unmixed stars, in the mean.
- lithium, which is a very fragile element (destroyed as soon as the temperature reaches 2 10⁶K) to be practically completely burnt.

Taking advantage of the high quality of the UVES spectra it was possible to measure not only the abundance of ¹²C but also the ¹³C abundance from the very weak lines of ¹³C (Fig. 1).

Figure 1: Spectrum of a very old star which has one thousand times less metals than the Sun (crosses) and the synthetic spectrum computed with different abundances of ¹³C (blue lines) . The weak ¹³C lines are clearly visible.
Click on the image to enlarge it

There is an excellent correlation between the ¹³C abundance and the nitrogen abundance. In Fig. 2 log (¹²C/¹³C) is plotted versus log (C/N). C/N is a very sensitive index of the transformation of carbon into nitrogen.
Two groups of stars are clearly separated. In the first group, both the C/N and the ¹²C/¹³C ratios are high (the abundance of ¹³C is low) while, in a second group, ("mixed" stars) a high quantity of carbon has been transformed into nitrogen (C/N and ¹²C/¹³C are low)

Figure 2: log (¹²C/¹³C) vs. log C/N. In the first group of star (in blue) C/N and ¹²C/¹³C are large and in the second group (in red) of evolved and mixed stars both C/N and ¹²C/¹³C are very weak.
Click on the image to enlarge it

Moreover, as expected, the C+N abundance in mixed and unmixed stars is about the same and that in the atmosphere of the mixed stars the lithium abundance is very low.
This "extra-mixing" seems to begin when, inside the star, the layer which burns hydrogen suddenly meets a hydrogen rich layer (this phase corresponds to the "bump" in the HR diagram).

To conclude...

Although standard models which take into account only convection, do not predict a mixing between the atmosphere of the red giant branch stars and the deep layers where carbon is transformed into nitrogen, this mixing is observed in the evolved group of very metal-poor giants.
This phenomenon affects dramatically the abundance of carbon and nitrogen (and even sometimes sodium) in the atmosphere of the stars and, as a consequence the abundance of these elements in the pristine gas of the Galaxy can be deduced only from less evolved "unmixed" old stars which have not undergone "extra-mixing".
On the other hand, nitrogen and carbon abundances in extremely metal-poor mixed stars bring strong constraints to the non-standard models. What is the cause of this extra-mixing: rotation ? magnetic fields? gravity waves ?

References :
First Stars VI - Abundances of C,N,O,Li, and mixing in extremely metal-poor giants. Galactic evolution of the light elements
Spite, Monique, Cayrel, Roger, Plez, Bertrand, Hill, Vanessa, Spite, Francois, Depagne Eric, François Patrick et al. 2005, Astronomy & Astrophysics, 430, 655

First Stars IX - Mixing in extremely metal-poor giants. Variation of the ¹²C/¹³C. ratio (Astronomy & Astrophysics, in press)
Spite, Monique, Cayrel, Roger, Hill, Vanessa, Spite, Francois, Plez, Bertrand, Bonifacio, Piercarlo et al.
http://fr.arxiv.org/abs/astro-ph/0605056

Contact
Monique Spite (Observatoire de Paris, GEPI)

Pristine abundances of C, N, and... mixing issues in the stellar interiors