Astron. Astrophys. 345, 233-243 (1999)

7. Discussion - possible scenarios of the formation of J-type carbon stars

ratios have been derived for 26 J-type carbon stars and the average is 4.7. The low ratios of J-type carbon stars have often been attributed to the mixing of CN-cycled material, since ratio is lowered to at the equilibrium of the CN-cycle. In other words, ratios can be , if the photospheres of J-type carbon stars consist of pure CN-cycled material. This is the case expected if the CN-cycle operates at the bottom of the convective envelope (Hot Bottom Burning, HBB). Massive stars, , develop deep convective envelopes with very high base temperature, leading to the operation of the CN-cycle (Sugimoto 1971, Iben 1975, Scalo et al. 1975, Renzini & Voli 1981, Blöcker & Schönberner 1991). However, the operation of the CN-cycle leads to the conversion of carbon into nitrogen, therefore, prevents stars from becoming carbon-rich. In fact, the HBB is a possible mechanism to keep massive stars from becoming carbon stars, and has been investigated as a way to interpret the absence of massive carbon stars in the Magellanic Clouds.

The extra mixing process suggested by Boothroyd et al. (1995) and Wasserburg et al. (1995) is of interest for understanding low ratios in J-type carbon stars. They introduce deep circulation currents below the bottom of the standard convective envelope. The bottom of the convective envelope remains cool, while the circulation currents mix material down to hot layers where the CN-cycle operates (cool bottom processing). Wasserburg et al. (1995) show that ratio is lowered to 4 in a 1 model, while C/O ratio exceeds 1 at a certain time on the AGB by addition of ¹²C synthesized in the thermal pulse. This prediction is consistent with the ratios we have derived here. But their model also predicts large nitrogen enrichments by a factor of 3 to 6. Lambert et al. (1986) determined nitrogen abundances from CN lines in the infrared region for four J-type carbon stars (RY Dra, T Lyr, VX And, and Y CVn), and their result shows that the nitrogen abundances in these stars are sub-solar. They conclude that the CN-cycle is inadequate for explaining low ratios in J-type carbon stars, while they also mention that RY Dra and Y CVn are relatively nitrogen-rich, and therefore that the operation of the CN-cycle might be possible. For our program stars, nitrogen abundances have not been determined, except for the four stars analyzed by Lambert et al. (1986). Because most of nitrogen atoms are locked up into N₂ molecules, it is not likely that the uncertainty of nitrogen abundance has a large effect on the temperature stratifications of the models or the resulting ratios. Nevertheless, it is highly desirable to determine nitrogen abundances in more J-type carbon stars in order to clarify the origin of low ratios and their evolutionary status.

A significant fraction of the stars studied here show ratios as low as , which are lower than the value at the equilibrium of the CN-cycle. Though the uncertainties of the ratios are relatively large for some of those stars, this result implies that those J-type carbon stars may have experienced non-equilibrium processes or the hot CNO-cycle on their way of evolution.

Another possible scenario of the formation of J-type stars is that they evolve from R-type carbon stars, which also have small ratios, 8 9 (Dominy 1984). A statistical parallax study (Vandervort 1958) shows that R-type carbon stars have 100 , which is the luminosity achieved by the He-core burning (Scalo 1976), while Dominy (1984) shows that their effective temperatures are as high as 4000-5000 K. Namely, their locations on the HR diagram correspond to those of the horizontal branch stars. It means that the change of the photospheric composition from oxygen-rich to carbon-rich should take place by the time they come to the horizontal branch. Dominy (1984) suggests that the mixing at the He-core flash could turn oxygen-rich atmospheres to carbon-rich. It will be interesting to perform theoretical calculations of the further evolution of R-type carbon stars, to examine whether the abundances and the isotope ratios observed in J-type carbon stars can be reproduced.

The formation of silicate carbon stars is also controversial. At present, the most plausible scenario may be the binary model proposed by Morris (1987, 1990), Lloyd-Evans (1990), and Lambert et al. (1990). The picture depicted by this scenario is that the material shed by an oxygen-rich primary star, possibly at the He-core flash, is stored in the accretion disk around a low mass companion until the primary becomes a carbon star. Barnbaum et al. (1991) show that the variations of radial velocities observed for EU And, BM Gem, and V 778 Cyg are consistent with motion in binary systems. Recently Kahane et al. (1998) detected a narrow feature of CO emission (J = 1-0 and 2-1) toward BM Gem, and they suggest that it is attributable to a circumbinary disk which is distorted or puffed-up. However, as Lambert et al. (1990) point out, a critical issue for this model is the stability of the accretion disk. Namely, if it is formed when the primary star is on the horizontal branch, it must survive until the primary becomes a luminous carbon star. It is impossible to verify this scenario based on our result of ratios alone. However, as the previous section shows, the ratios in the five silicate carbon stars studied in the present work show no peculiar values as compared with those of other J-type carbon stars. This implies that the mechanism responsible for low ratios in silicate carbon stars might be the same with that working in other J-type carbon stars. It might be inferred that the progenitors of silicate carbon stars and other J-type carbon stars are the same, R-type carbon stars for example, and that some descendants could be observed as silicate carbon stars when the conditions for the stability of the accretion disk, such as the mass ratio of the primary to the companion, the separation, etc., are met.

© European Southern Observatory (ESO) 1999

Online publication: April 12, 1999
helpdesk.link@springer.de