next previous
Up: Stellar and circumstellar evolution


Subsections

   
2 Chemical evolution

It is well known that K and IRAS absolute magnitudes reflect the properties of various parts of the star. The K magnitude depends mainly on the characteristics of the stellar surface, whereas IRAS fluxes are linked to envelope thickness and dust composition. Using both types of magnitudes, information about the stellar and circumstellar properties can be obtained.

Here we mainly use the individual absolute magnitudes and assigned galactic populations estimated in Paper I. It is convenient to recall that our sample was found to be representative of the LPV population as far as the kinematics and the brightest luminosities are concerned, but is under-representative for K and IRAS faint stars (see Paper I).

   
2.1 From O-rich to C-rich LPVs


  \begin{figure}
\includegraphics[width=12cm,clip]{1429f1.eps}\par
\end{figure} Figure 1: Distribution of the individual estimated K luminosities and 12-25 IRAS colors according to the assigned kinematical groups and to spectral types.

Figure 1 shows the distribution of the individual estimated K absolute magnitudes and 25-12[*] indices deduced from the estimated IRAS absolute magnitudes according to the assigned kinematical groups and spectral types. A bimodal distribution of stars with a deficit in the number of stars around 25-12=-0.2 is clear, mainly for the lower mass stars (disk 2 and old disk). Section 3 will examine in detail this gap for O-rich stars. In the present section we focus on the fact that this area contains mainly C-rich stars belonging to disk population. Moreover it corresponds to the range of the (25-12) index with the lowest ratio of known variable stars among the IRAS sources with colors similar to LPVs colors (Paper I, Fig. 4).

This gap has already been reported by Habing (1989) and interpreted as the separation between stars with and without a circumstellar envelope. The deficit is very marked in regions I and II of Habing's IRAS color-color diagram[*], which correspond to non-variable and variable (mainly O-rich) stars, and slighter in region VII, which contains C-rich variable stars, in agreement with our results. Thus, stars with a thin circumstellar envelope (f) are clustered around 25-12=0 and almost all of them are O-rich LPVs. However, the kinematical study allowed us to assess the differences according to the initial mass of the stars.

The most massive stars can evolve to C-rich LPVs after a number of dredge-ups that enrich the external shells of the star in carbon. In these stars the C/O ratio becomes larger than 1 and when it is around 1, the star is an S star. At the same time, strong changes take place in the circumstellar envelope, which becomes dominated by C-rich grains. The 25-12 index increases in conformity with the loop in the IRAS (25-12, 60-25) color-color diagram predicted by Willems & de Jong (1988) and calculated by Chan & Kwok (1988). The areas occupied by O-rich and C-rich disk LPVs in Fig. 1 reflect this loop. Indeed the 25-12 index decreases from stars with a thin circumstellar envelope (f) to O-rich LPVs with a thick envelope (b) and increases again for C LPVs.

However, our luminosity calibrations suggest that the phenomenon does not strongly induce a difference in the K distributions of O-rich and C-rich disk LPVs. Moreover, the (K, IRAS) luminosity diagrams of Figs. 2 and 3 show that the loop is caused by both the decreasing 12 and 25 luminosities when the star becomes C-rich, but that this decrease is stronger in the 25 filter.

C-rich irregular and SRb stars mainly belong to the disk 1 population (Paper I, Sect. 6.4). They are mainly located at the upper end of the AGB. Therefore, the change from a more regular variable star (Mira) to a less regular one (L or SRb) may be associated with an increase in the non-linear behaviour of the massive pulsating LPVs due to interactions of the pulsation phenomenon and a very thick and dynamically unstable envelope.

   
2.2 Peculiar evolutive phases

The above global evolutive scenario along the AGB can be refined by examining stars which correspond to peculiar short-lived stages during which the star fundamentally changes (S or Tc stars), or to one of the most advanced stages of evolution (OH emitters).

   
2.2.1 Tc stars

According to the models, the s-elements processed in a star can be brought to the surface by convective dredge-ups. When a sufficient quantity of s-elements has been brought to its surface, an O-rich star becomes an S star. S stars present a C/O ratio close to 1 and are generally considered as a transition phase between O-rich and C-rich stars.

Some S stars are enriched in Tc, indicating that this material was brought to the surface by recent (in the last few million years) dredge-ups, as modeled by Mowlavi (1998) in agreement with the results obtained by Van Eck et al. (1998) who compared Tc and no-Tc S stars. Our list of Tc-rich S LPVs is taken from Van Eck's thesis (1999).

On another hand some O-rich LPVs (i.e. LPVs classified with an M spectral type) can also be enriched in Tc. Such peculiar stars were studied by Little et al. (1987) and they offer great potential as a possible constraint on the modelization of dredge-ups. Table 1 shows the individual estimated absolute magnitudes and the assigned group of these Tc stars.

   
Table 1: Individual K, 12 and 25 luminosities, with assigned crossing K (IRAS) group of Tc O-rich and S spectral type LPVs. Variability ( ${\rm M}={\rm Mira}$, ${\rm SR}={\rm semi}$ regular, ${\rm L}={\rm irregular}$) types and possible specificity ( ${\rm Tc}={\rm Technetium star}$, ${\rm BD}={\rm bright}$ galactic disk star) are given.
HIP id name types group K 12 25 pec
8 Z Peg MO ODb -6.69 -8.90 -9.58 Tc
1236 S Scl MO ODb -7.11 -9.08 -9.63 Tc
77615 R Ser MO ODb -6.97 -9.57 -10.17 Tc
90493 RV Sgr MO D1 -6.92 -8.43 -9.06 Tc
104451 T Cep MO D1 -8.08 -9.99 -10.60 Tc
110736 S Gru MO D2b -7.48 -9.98 -10.71 Tc
                 
1901 R And MS ODb -6.79 -9.69 -10.54 Tc
10687 W And MS D1 -8.83 -11.95 -12.65 Tc, BD
34356 R Gem MS D1 -7.69 -9.81 -10.36 Tc
35045 AA Cam LS ODb -6.91 -8.19 -8.84 Tc
38502 NQ Pup LS D2f -6.73 -6.97 -7.05 Tc
65835 R Hya MS D2b -8.52 -11.04 -11.60 Tc
87850 OP Her SRS D2b -7.30 -8.30 -8.65 Tc
94706 T Sgr MS D2b -7.95 -10.30 -10.94 Tc
97629 khi Cyg MS D1 -7.48 -9.71 -9.88 Tc
98856 AA Cyg SRS D2b -8.83 -10.66 -11.35 Tc
110478 pi.1 Gru SRS D1 -8.16 -9.82 -10.59 Tc
113131 HR Peg SRS D1 -6.78 -7.39 -7.67 Tc
                 
17296 BD Cam LS D2f -6.01 -6.53 -6.63  
33824 R Lyn MS ODf -7.82 -7.85 -8.03  
40977 V Cnc MS D2f -6.47 -6.85 -6.93  
101270 AD Cyg LS D1 -9.18 -10.49 -10.88 BD
110146 X Aqr MS ODb -6.68 -8.92 -9.57  
112784 SX Peg MS D2f -6.28 -7.00 -7.11  
114347 GZ Peg SRSa D1 -7.63 -8.45 -8.61  

We would like to highlight the correlation between location in the plane (12, 25) of Tc LPVs and the limit between O-rich and C-rich regions, regardless of their S or M spectral type (see Fig. 4). The only exception is R And, which will be discussed later in this paper. However no differences are found between Tc O-rich and Tc S LPVs. Tc O-rich LPVs have the same 12 and 25 luminosities as O-rich stars. They are probably LPVs enriched in Tc by a recent dredge-up, but not efficient enough either to make the C/O ratio close to 1 or to drastically alter the circumstellar envelope. Tc S LPVs are mainly assigned to disk population (10/12), which is not valid for Tc O-rich LPVs. This suggests that the dredge-up is more efficient in changing the C/O surface ratio up to 1 for more massive stars. No definitive conclusions can be reached owing to the scarcity of Tc LPVs in the sample.

It is also important to note that all these stars are more luminous than $\simeq -6.5$ mag in K, in agreement with Van Eck et al. (1998), the bolometric correction for this type of stars being around 3 mag. This confirms the predicted location of the first thermal pulse (Mowlavi 1998) and the quite early operative third dredge-up on the TP-AGB (Van Eck 1999).

Finally, some individual Tc stars in our sample have specific properties that require a specific discussion:

   
2.2.2 R Hor and Tc enrichment

R Hor was found to be Tc enriched by Little et al. (1987) but not confirmed as such by Van Eck (1999) and it is the only Tc star assigned to the extended disk population. A Bayesian classification process can lead to some misclassification, but R Hor is at the limit of the O-rich LPVs area, close to two C-rich LPVs (RS Lup and V CrB) (see Table 3). Thus, its Tc enrichment is questionable.

Another point is the way of enrichment in Tc and in C of such a deficient and low mass star. V CrB was reported as probably "metal poor'' by Hron et al. (1998) from ISO data. This agrees with our assignation to ED (Table 3), which seems a priori doubtful for a carbon star.

If this is confirmed, it would be an interesting constraint to evolutive models.

   
2.2.3 Non-Tc S-type stars

Some of the non-Tc S stars in our sample can be extrinsic S stars. These stars are not enriched in s-elements by internal nucleosynthesis and dredge-ups but by mass transfer from a more evolved companion. This is probably the case of X Aqr, BD Cam, V Cnc and SX Peg, for which a duplicity flag is given in the HIPPARCOS catalog. Except for X Aqr, they are in Fig. 4 at the lower limit in K of the old disk population with thin envelope (ODf), to which they are assigned. Moreover, the four of them have a K luminosity under -6.5 mag i.e. below the threshold of thermal pulses on AGB, confirming the Van Eck's (1999) result.

There are three other non-Tc S stars in our sample. They are the least luminous in 12 and 25 bands and have the largest 25-12 index among stars with their K absolute magnitudes. AD Cyg is assigned to the bright disk and it is more luminous than all C stars. It is a massive star and we can assume that its evolution along the AGB is very rapid. On the other hand, in the (K, 25-12), (K, 12), and (K, 25) planes, R Lyn and GZ Peg are close to the line on which AD Cyg and three extrinsic S stars assigned to disk 2 population with a thin envelope (D2f) (BD Cam, V Cnc, SX Peg) are located. These locations seem to confirm the non-Tc character of these stars, which are probably extrinsic S stars. However, given the difficulties in detecting duplicity, we failed to confirm this result.

Although the number of extrinsic S stars in our sample prevented us from reaching any definitive conclusion, these results suggest that an extrinsic S enrichment can accelerate the evolution along the AGB with formation of a circumstellar envelope closer to a carbon than to a silicated composition, before any enrichment by the star's own nucleosysthesis and dredge-ups.

   
2.2.4 OH stars

Some O-rich LPVs are maser emitters on the radio frequencies corresponding to OH bands, the masers being pumped by infrared photons.

These stars emit at the principal frequencies of the main transition (1665/67 MHz) and some also emit at 1612 MHz. The star is classified as OHII or OHI according to a stronger or a lower emission at the secondary frequency with respect to the main band. All of them are O-rich LPVs.

A systematic research of OH masers for stars in the solar neighbourhood has been carried out by Sivagnanam et al. (1989, 1990), Lewis et al. (1995) and Szymczack et al. (1995).

   
Table 2: Individual K, 12 and 25 luminosities, with assigned crossing K (IRAS) group of OH maser emitters Mira.
HIP id name types group K 12 25 pec
11350 R Cet MO ODb -6.69 -9.29 -10.34 OHI
19567 W Eri MO D2b -6.26 -9.48 -10.21 OHI
21766 R Cae MO D1 -8.56 -10.80 -11.48 OHI
25673 S Ori MO D2b -7.91 -10.20 -10.94 OHI
26675 RU Aur MO ODb -6.90 -10.23 -11.09 OHI
27286 S Col MO D2b -8.04 -10.09 -10.87 OHI
36669 Z Pup MO D1 -8.17 -11.05 -11.86 OHI
40534 R Cnc MO D2b -8.61 -10.83 -11.55 OHI
47066 X Hya MO ODb -6.48 -8.52 -9.08 OHI
47886 R LMi MO D1 -8.41 -10.73 -11.37 OHI
48036 R Leo MO D1 -7.67 -9.91 -10.18 OHI
58854 R Com MO D1 -6.80 -8.41 -9.22 OHI
67626 RX Cen MO ODb -6.80 -9.12 -9.90 OHI
69346 RU Hya MO ODb -6.49 -8.95 -9.76 OHII
69816 U UMi MO D1 -6.98 -8.58 -9.18 OHI
70669 RS Vir MO D1 -7.30 -10.27 -11.18 OHII
74350 Y Lib MO ODb -6.15 -8.11 -8.94 OHI
75143 S CrB MO D2b -8.01 -10.43 -11.45 OHII
75170 S Ser MO D1 -6.72 -8.76 -9.41 OHI
79233 RU Her MO D1 -7.93 -10.07 -10.93 OHI
80488 U Her MO D2b -7.95 -11.06 -11.64 OHII
91389 X Oph MO D1 -8.16 -11.11 -11.69 OHI
93820 R Aql MO ODb -7.31 -9.48 -10.49 OHII
98077 RR Sgr MO D2b -7.64 -10.36 -10.97 OHI
98220 RR Aql MO ODb -6.88 -9.72 -10.48 OHII
114114 R Peg MO D2b -7.35 -10.12 -10.79 OHI

Individual estimated absolute magnitudes and assigned groups of OH stars in our sample are given in Table 2. Figure 5 shows their location in the distributions of the various absolute magnitudes. The distinction between OHI and OHII is not useful for our purposes because no difference was found between them in our analysis.

As expected, no OH star was found with a thin envelope, i.e. assigned to an f group. At a given K, they are the brightest in the 12 and 25 bands and all (except R Leo) have a 25-12 index corresponding to a thick circumstellar envelope. Their K luminosity distribution is the same as that of non-OH O-rich LPVs. However, our previous V calibration (Mennessier et al. 1999) indicates the extent to which the presence of a thick envelope induces an absorption in the visible range and confirms that after a new growth of the envelope the star becomes fainter and fainter in the visible range and turns into an OH-IR source. All these results agree with the current model of OH sources: a maser emission pumped by photons of an infrared thick envelope that depends on the mass of the star and the mass-loss ratio.

Finally, the kinematic assignation of OH emitters also provides information about these stars. Ten stars were assigned to the disk 1 population and they can thus be considered LPVs with massive progenitors. However we found that this population group is rather attractive for C-rich stars and repulsive for O-rich stars (Paper I). These ten OH emitters have a K luminosity brighter than about -8 mag. They are located in the same area as the Tc S LPVs in the (K, 25-12), (K, 12) and (K, 25) planes. Thus such stars probably have a mass at the limit of the capability to be sufficiently enriched in carbon by successive dredge-ups. As discussed in Sect. (4.2) some of them could be Hot Bottom Burning candidates.

The other OH sources of the sample, assigned to the disk 2 or old disk population, are less massive.


next previous
Up: Stellar and circumstellar evolution

Copyright ESO 2001