Astron. Astrophys. 342, 831-838 (1999)

4. Discussion

Since the discovery of the Li-rich K giants several mechanisms to explain the overabundance of Li have been suggested (Wallerstein and Sneden 1982, Gratton and D'Antona 1989, Brown et al. 1989, Pilachowski et al. 1990, de la Reza et al. 1996, 1997). Engulfing orbiting planets and/or brown dwarfs, binarity and mass loss were some of the suggestions to account for the overabundance in Li and far-infrared excess. A model scenario has recently been introduced by de la Reza and colleagues. In this scenario all ordinary, Li-poor, K giants become Li-rich during a short time (10⁵ years) when compared to the red giant phase of years. In this "Li-period" a large number of the G and K giants are associated with an expanding thin circumstellar shell supposedly triggered by an abrupt internal mixing mechanism resulting in a surface Li enrichment.

4.1. Absolute magnitudes

In order to study the position of Li-rich and Li-poor (normal) red giants in the H-R diagram we have determined the absolute visual magnitudes M_V of all the well studied Li-rich K giants from their Hipparcos parallaxes (Table 3).

Table 3. Lithium rich K giant stars with data taken from literature. Log and 12 C  / 13 C values can be found in da Silva et al. (1995). Rotational velocities are coming from de Medeiros et al. (1996), except for HD 120602 and HD 121710 (Fekel & Balachandran 1993) and HD 233517 (Fekel et al. 1996).

Many of these are bright and relatively nearby stars and therefore the interstellar reddening is not too large. Using the B-V colors and spectral types we derived the E(B-V) values. The M_V values given in Table 3 are corrected for the extinction. The uncertainty in M_V values as a result of extinction corrections are of the order of 0.25 magnitude. We have also determined the M_V values of our program stars from their Hipparcos parallaxes (Table 1). Fig. 2 shows the position of the Li-rich red giants (Table 3) and a few normal red giants (Table 1) in the H-R diagram together with the evolutionary tracks taken from Schröder (1998) and adapted from Schröder et al. (1998).

Fig. 2. The position of Li-rich giants in the H-R diagram. The absolute visual magnitudes (Mv) of Li-rich giants are calculated from their Hipparcos parallaxes. symbols are stars observed in this paper (Table 1); symbols are Lithium-rich K giants stars from Table 3. The position of "clump giants" (indicated by a rectangle) and the evolutionary tracks of giants from 1 to 3.2 solar masses are adapted from Fig. 1 in Schröder et al. (1998).

From Fig. 2 it seems that Li-rich giants have a range of masses and luminosities. Most of them are more luminous than the "clump- giants". HD 205349 was found to be a Li-rich (log (Li) = 1.9) supergiant by Brown et al. (1989). The Hipparcos parallax yield an absolute magnitude M_V = -3.43 (Table 3) confirming the supergiant status of this star. From our analysis we find HD 152786 also to be supergiant with M_V = -3.95 and log (Li) = 1.3. Fig. 2 suggests that whatever may be the mechanism for the overabundance of Li, it is not confined to a narrow range of luminosities and masses of giants. Two Li-rich giants HD 19745 (log (Li) = 4.75) and HD 95799 (log (Li) = 3.22) are not in the Hipparcos catalogue. For HD 19745 we adopt the absolute magnitude 0.42 given by de la Reza & da Silva (1995), and for HD 95799 we estimated the M_V value in assuming the star as a member of the NGC 3532 open cluster. Fig. 2 clearly shows that the Li-rich giants are not pre-main-sequence stars. Some authors have speculated that these are Li-rich because they are pre-main-sequence stars.

From Fig. 3 we can explore a possible connection between Li abundance and absolute magnitude. There seems to be a weak correlation between Li abundance and absolute magnitude, which indicates that among the Li-rich giants, relatively less luminous giants are more Li-rich. This result indicates that the low mass giants which spend long time on the Red Giant Branch (RGB) may produce more Li as they are expected to undergo several episodes mixing. However we need to find more Li-rich giants and derive their luminosities in order to fully explore the relation between Li abundance and luminosity among the Li-rich giants.

Fig. 3. The Li abundance of Li-rich giants plotted against their absolute visual magnitudes. symbols are stars observed in this paper (Table 2); symbols are Lithium-rich K giants stars from Table 3.

4.2. 12 C /13 C

da Silva et al. (1995) discussed the carbon isotope ratios in Li-rich giants. There is no clear correlation between Li abundance and 12 C /13 C ratio. They found that the three most Li-rich giants HD 19745 (log (Li) = 4.08 (LTE) or 4.75 (NLTE)), HD 39853 (log (Li) = 2.9 or 3.9) and HD 95799 (log (Li) = 3.22) to show 12 C /13 C 15, indicating extra mixing. The 12 C /13 C ratio in the Li-rich giants ranges from 6 to 28 (Table 3, Fig. 4) similar to that found for normal K-giants (Gilroy 1989). There seems to be no clear relation between the 12 C /13 C ratio and absolute visual magnitude (Fig. 4). The position of Li-rich giants in the H-R diagram and their carbon isotope ratio indicate that they are evolved and have experienced normal amount of convective mixing. Some of them with low carbon-isotope ratio may have gone through additional mixing on the giant branch.

Fig. 4. The carbon isotope ratios of Li-rich giants plotted against their absolute visual magnitudes. symbols are Lithium-rich K giants stars from Table 3.

4.3. Rotational velocities

de Medeiros et al. (1996) determined precise rotational velocities of several Li-rich giants (Table 3) using CORAVEL spectrometer. Except for a few, the Li-rich giants show normal rotational velocities with respect to typical Li-normal giants of the same spectral type. de Medeiros et al. (1996) found no indication of binarity for the Li-rich giants. Fig. 5 shows that there is no correlation between Li abundance and rotation for the Li-rich giants. There are few rapidly rotating Li-rich giants (Table 3) (for example HD 9746, HD 219025 and HDE 233517). Fekel and Balachandran (1993) suggest a connection between rapid rotation and high Li abundance in giants. They suggested a scenario in which the surface convection zone reaches the rapidly rotating core as a star begins its first ascent of the giant branch and dredges up to the surface high angular momentum material and freshly synthesized Li. Fekel et al. (1996) suggest that the processes causing rapid rotation, large Li abundance, and infrared excess are triggered at the base of the giant branch when the convection zone reaches the rapidly rotating core of low-mass stars. This explanation is not able to account for the slowly rotating Li-rich giants.

Fig. 5. The Li abundances of Li-rich giants plotted against their rotational velocities (same symbols as Fig. 3).

Finally most of the Li-rich giants show moderate or weak chromospheric activity (de la Reza and da Silva 1995, Fekel and Balachandran 1993) indicating no correlation between stellar activity and Li-abundance (Randich et al. 1993, 1994)

4.4. Infrared excess

The dust shells around some of the Li-rich giants are cold and probably detached, though detachment has not yet been proved. Whitelock et al. (1991) and Zuckerman (1993) have suggested that binarity is implicated when large amounts of dust are found near evolved stars. There is no evidence that these stars are preferentially members of binary systems. Among the stars analyzed in this work, HD 169689 is known as an eclipsing binary of Algol type (period of 385 days), and HD 176884 is a member of a visual binary. Besides, HD 9746 = OP And is a chromospherically active star with a rotational period of 2.36 days (ESA HIPPARCOS catalogue) and is not known as a binary. Further investigation on binarity among G-K giants with infrared excess is in progress with the spectrovelocimeter CORAVEL and will be published soon.

Some giants like HDE 233517 and HD 219025 are very Li-rich, have high rotational velocities and also have large infrared excesses but others have excess Li and no far-infrared excess and vice versa. To account for the circumstellar dust around the late-type giants evaporation of planets, brown-dwarfs, and cometary belts are suggested, as the star evolves from the main-sequence phase to a giant (Zuckerman et al. 1995, Plets et al. 1997). To account for the overabundance of Li also such scenarios were proposed (Brown et al. 1989, Gratton and D'Antona 1989). The presence of Li-rich giants with no circumstellar dust, with no enhanced chromospheric activity and with slow rotation indicates that the Li-enrichment may not be linked to any of these parameters.

4.5. Creation of Li

Recently Sackmann and Boothroyd (1998) demonstrated that Li can be created in low mass red giant stars via the Cameron- Fowler mechanism, due to extra deep mixing and the associated "cool bottom processing". They conclude that the amount of Li produced can exceed log (Li) = 4 but depends critically on the details of the extra mixing mechanism (mixing speed, geometry, episodicity). Sackmann and Boothroyd (1998) predicted that if the deep circulation is a relatively long-lived, continuous process, lithium-rich red giants should be completely devoided of beryllium and boron. In this context determination of Be abundance in the Li-rich red giants (Table 3) is important. The study of Sackmann and Boothroyd (1998) is able to explain the overabundance Li in the red giants. However, they have not considered the rapid rotation and mass loss process during this deep circulation. It is not yet clear whether this deep circulation and cool bottom processing triggers mass loss. Also the effect of rapid rotation on deep circulation and cool bottom processing needs to be explored in order to understand the Li-rich red giants with high rotational velocities and circumstellar dust.

© European Southern Observatory (ESO) 1999

Online publication: February 23, 1999
helpdesk.link@springer.de