Planificación

Depending on which element - oxygen or carbon - is most abundant, different molecu- lar absorption bands in the visual and infrared (e.g. Lançon & Wood 2000) dominate the spectrum of M- and C-type stars. Molecular equilibrium calculations given by Russell (1934), were the first quantitative explanation of the important differences in the spectrum of oxygen- and carbon-rich stars. The strong binding energy of the CO compound regu- lates the molecular abundances for different C/O ratios. While strong TiO, VO and H2O bands are present for C/O<1, the C/O> 1 case shows C2, CN and SiC absorption fea- tures. Furthermore, this carbon-enhancement also affects the chemical composition of the dust grains as amorphous silicates and amorphous carbon grains are respectively present for each case (e.g. Ivezic & Elitzur 1995, Habing 1996). Observationally, these spectral differences appear as a sharp discontinuity in near- and far-infrared colors between M and C-type stars. Data is being released in huge amounts (e.g. the DENIS and 2MASS projects) and theory should account for these observations.

Tsuji (1966) had already shown the importance of different molecular opacity sources and was the first to include the effects of absorbers such as H2O, CO and OH in opacity calculations. But in stellar evolution codes low-temperature molecular opacities are still incompletely treated since the adopted descriptions (Alexander & Ferguson 1994, Fergu- son et al. 2005) correspond to gas mixtures for a single value of the C/O ratio. Only few works have proposed molecular opacity tables for variable C/O ratios (Alexander et al. 1983, Lucy et al. 1986) and some (Bessell et al. 1989, 1991) have presented analytical fits, as a function of carbon and oxygen abundances, to some already existing tables (Alexander et al. 1983). But even fewer works have coupled such opacities to evolutionary calculations (Scalo & Ulrich 1975).

4.2 TP-AGB Evolution with Carbon-Enhancement

With an important publication (Marigo 2002) on the effects of molecular opacities for varying surface C/O ratios, P. Marigo is the one that has most recently presented work in which synthetic AGB models are connected to an appropriate opacity description at low temperatures. The adopted procedure to compute the molecular opacities, through analytical fit relations, closely resembles that of Scalo & Ulrich (1975) and is incorporated in the P. Marigo synthetic code for TP-AGB evolution. The possibility to consistently compute the opacities for any chemical composition, during the evolutionary calculations is a huge advantage of this approach and the effects detected in the models help account for a number of observational properties of carbon stars.

The molecular opacity routine shows an abrupt change in the dominant opacity sources at the transition (C/O= 1) from oxygen- to carbon-rich envelopes. The most significant consequence of this effect is the sudden cooling of the stellar models (see section 7.3). This is important because, as the mass-loss descriptions depends on the effective temperature, the decrease in temperature leads to a increase in mass-loss when the star becomes a carbon star. The low temperatures and high mass-loss rates have many direct implications on AGB evolution. They can alter the expected number of C with respect to M and S stars, the total number of TPs, the number of TPs during the carbon-rich phase, the nucleosynthesis, the maximum C/O ratio obtained etc. For example, because mass-loss is higher during the carbon-rich phase, TP-AGB evolution is shorter, fewer TPs occur and therefore less nucleosynthesis.

Nevertheless the adopted simplifications in the Marigo approach, constitute a clear lim- itation of the opacity calculations (in section 7.3.3 we will show how these opacities are coupled to our AGB models). The total Rosseland mean opacity is simply taken to be the sum of the Rosseland mean opacity of each individual molecule. This is supposed to hold under most physical conditions met in AGB envelopes, even if it is not generally correct. Furthermore, by neglecting TiO and VO molecules for oxygen-rich compositions as well as HCN and C2H2for carbon-rich ones, only a limited number of molecular species - they are however among the most relevant opacity sources in AGB stars - are included. Finally, the AGB models to which these opacities were coupled are synthetic, and therefore simply represent an approximation to detailed stellar evolution calculations.

On one hand, multidimensional interpolations between pre-tabulated values of detailed opacity calculations - like the ones of Iglesias & Rogers (1996), Alexander & Ferguson (1994) and Ferguson et al. (2005) - would be a more accurate alternative. In that case, the parameters that should at least be taken into account are density, temperature, total metal content and hydrogen, carbon, oxygen and nitrogen abundances. On the other hand, there is a necessity to couple these opacities to a stellar code which actually solves the entire set of stellar structure and nuclear generation equations in order to get a fully consistent model directly parameterizing the occurring physical processes.

A large enough grid of tables, to adequately cover the abundance changes found in AGB envelopes, with accurate molecular opacities was not available until now.

In this context, J. Ferguson from the Wichita State University (WSU) low-temperature opacity group, produced detailed opacity tables (Ferguson 2006) for different C/O ratio compositions, tailored to the specific needs of TP-AGB models produced by the Garching

detailed stellar evolution code.

4.3. New Low-Temperature Opacities for Variable Chemical