The validity of the hypothesis of LTE questioned from the observational side

Therefore, from the theoretical point of view mentioned above, there should be no need to worry about the NLTE overionization effect for Fe (and presumably also for other Fe-group elements) in the metal-rich or near-Solar metallicity stars of our present concern. Yet, a few groups have recently reported that appreciable inconsistencies were observed in the abundances obtained under the assumption of LTE even for disk stars of around solar metallicity.

For example, Bodhagee et al. (2003) and Gilli et al. (2006) found in their analysis of planet-host and comparison stars of early K through late F types that Fe-normalized abundances [X/Fe] derived from lines of several neutral elements (e.g. Ti i, V i, Mn i, Co i) exhibit systematic tendencies ([X/Fe] increasing with decreasing Teff) over a wide Teff range from ~5,000K to ~6,000 K, which they attributed to some NLTE effect. We feel, however, that such an apparent Teff dependence as they claimed might be due to some spurious effect (e.g. blending with unknown lines whose importance would progressively increase toward a lower Teff), because we could not confirm such a tendency in our recent analysis (Takeda 2007) of 160 F, G, and K stars including 27 planet-host stars similar to theirs, as

5,000

5,000 6,000 7,000

6,000 7,000

5,000 6,000 7,000

5,000

m

il

- (d)

6,000

Figure 32.4. Plots of [Ti/Fe], [V/Fe], [Mn/Fe], and [Co/Fe] with TeS, constructed from the results of Takeda's (2007) analysis of 160 F, G, and K stars. The zero levels are indicated by dashed lines. Open and filled symbols denote planet-host stars and non-planet-host stars, respectively.

4,500 5,000 5,500

4,500 5,000 5,500

6,000

Figure 32.4. Plots of [Ti/Fe], [V/Fe], [Mn/Fe], and [Co/Fe] with TeS, constructed from the results of Takeda's (2007) analysis of 160 F, G, and K stars. The zero levels are indicated by dashed lines. Open and filled symbols denote planet-host stars and non-planet-host stars, respectively.

Figure 32.5. The Fe abundances of Hyades G-K dwarfs obtained by Schuler et al. (2006b), re-plotted by ourselves as functions of Teff using the data presented in Table 5 in their paper. Abundances from Fe i and Fe ii lines are indicated by filled and open symbols, respectively. Together with the results for dwarfs (circles), those for three giants are also shown by triangles.

shown in Figure 32.4. In any case, further investigation of this matter should be carried out.

Another item of evidence for a definite breakdown of LTE has recently emerged from analyses of G-K dwarfs in Hyades. Although almost the same abundances should in principle be derived for such cluster members irrespective of Teff, Yong et al. (2004) reported a general tendency of A(Fe i) < A(Fe ii) together with a manifest rise (toward a lower Teff) of A(Fe ii) at Teff < 5,000 K. A similar result was obtained also by Schuler et al. (2006b) as shown in Figure 32.5, which we have redrawn using their data in order to facilitate comparison with Figure 4 of Yong et al. (2004). This result strongly suggests that NLTE overionization prevails for Hyades K dwarfs of Tef < 5,000 K, which raises the Fe ii population (while the concomitant reduction of the Fe i population is inconspicuous because Fe i is the dominant fraction in K dwarfs).

Meanwhile, we should clarify the cause of such an overionization in K dwarfs' atmospheres, since it cannot be explained within the framework of the classical model atmosphere, which predicts that UV radiation fields are essentially

Ti I

thermalized due to the enormously enhanced (atomic as well as molecular) line opacities. We suspect that the key may be the chromospheric activity. Schuler et al. (2006a, 2006b) found that the oxygen abundances of Hyades dwarfs derived from the infrared O i 7773 triplet also undergo a conspicuous rise with decreasing Teff at Teff < 5,000K (cf. Schuler et al. 2006b, Figure 10), just like the trend of Fe ii abundances. That is, neither the O i 7773 triplet nor Fe ii lines can yield reliable abundances (i.e. values are significantly overestimated) under the assumption of LTE in Hyades K dwarfs. Recalling here that the strength of the O i 7773 triplet is quite sensitive to the chromospheric temperature rise (e.g. Takeda 1995), we may speculate as a possibility that Hyades K dwarfs become progressively more active toward a lower Teff, and that the enhanced chromospheric UV radiation is the cause for the NLTE overionization.

If this is the case, such an appreciable NLTE effect in K stars might not simply apply to any stars in general; i.e. it may be comparatively younger stars that are mainly affected here, since the activity is closely related to the stellar age through a secular breaking of rotation due to loss of angular momentum. Actually, the Fe overionization forthe case of M34 (age ~2 x 108 yr), which is younger than Hyades (age ~8 x 108 yr), appears to be noticeable already at Teff ~ 5,500 K (cf. Schuler etal. 2003, Figure 5), which is earlier than in the case of Hyades (at Teff ~ 5,000 K). We hence suspect that this kind of NLTE effect in Fe may be less significant for most of the disk stars in general (whose ages are typically from ~109 to ~1010 yr). In this connection, it would be very interesting to study G-K dwarfs of old open clusters (such as M67, which is as old as the Sun).

Was this article helpful?

0 0

Post a comment