Nearby OB stars

Here we consider results for nearby massive stars, by which we mean OB stars within approximately 500 pc of the Sun. Within this distance we expect that the abundance dispersion due to the expected Galactic abundance gradient (~ 0.05 dex) can be neglected. We further restrict ourselves to results from NLTE calculations only, since we wish to compare the absolute abundances with Solar. Unfortunately most such results are based on NLTE line-formation calculations in which LTE line-blanketed model atmospheres have been adopted. There are currently very few results from full NLTE line-blanketed model atmospheres, but see Hunter et al. (2007). There have been essentially three studies matching these requirements; those of Kilian (1992), hereafter K92, Gies & Lambert (1992), GL92, and Daflon et al. (1999,2001), D99. These samples have been filtered further to remove super-giants (from GL92) and the mean values of [O/H] are listed in Table 6.1.

Referring to Table 6.1, one can see that the various NLTE results for nearby B-type stars are rather similar, while the left-hand panels of Figure 6.1 confirm that there are small systematic differences between the various sets of results. Indeed, for five stars in common between K92 and GL92 one finds a mean difference in [O/H] of 0.10 dex, the K92 results being the lower ones. This is a substantial part of the 0.15-dex difference between the complete samples. Furthermore, when one considers the trend of [O/H] with effective temperature (right-hand panels of Figure 6.1) it is clear that both GL92 and D99 display strong trends, though in opposite senses; in GL92 [O/H] increases with effective temperature whereas in D99 [O/H] decreases with effective temperature. It is only in the latter two cases that one finds mildly 'metal-rich' stars, but, since these lie at the extremes of the temperature ranges, it is clear that their metallicities must be suspect. The reason for these trends with

Stellar Images

25 30 35 Teff(kK)

25 30 35 Teff(kK)

25 30 35

25 30 35

20 25 30 35 7"eff(W0

Figure 6.1. Oxygen abundance ([O/H]) of B-type stars in the Solar neighbourhood. Top panel: results from Kilian (1992) (K92) using NLTE line formation and LTE models with partial line-blanketing, mean [O/H] = 8.54 dex. Central panels: results from Gies & Lambert (1992; GL92) using NLTE line formation and LTE models with line-blanketing, mean [O/H] = 8.69 dex. Bottom panels: results fromDaflon et al. (1999, 2001) (D99) using NLTE line formation and LTE line-blanketed models, mean [O/H] = 8.61 dex. Note the temperature trends in GL92 and D99; the trend in GL92 is further enhanced when one removes the two points above 25 000 K with the lowest [O/H], these being for a variable star and an SB2.

20 25 30 35 7"eff(W0

Figure 6.1. Oxygen abundance ([O/H]) of B-type stars in the Solar neighbourhood. Top panel: results from Kilian (1992) (K92) using NLTE line formation and LTE models with partial line-blanketing, mean [O/H] = 8.54 dex. Central panels: results from Gies & Lambert (1992; GL92) using NLTE line formation and LTE models with line-blanketing, mean [O/H] = 8.69 dex. Bottom panels: results fromDaflon et al. (1999, 2001) (D99) using NLTE line formation and LTE line-blanketed models, mean [O/H] = 8.61 dex. Note the temperature trends in GL92 and D99; the trend in GL92 is further enhanced when one removes the two points above 25 000 K with the lowest [O/H], these being for a variable star and an SB2.

temperature probably lies in the methods used to derive effective temperatures; K92 used the silicon-ionization-balance method, GL92 used the Stromgren [ci] index with empirical corrections to bring the results into line with the empirical estimates of Code et al. (1972) and D99 used another photometric method based on the Q-value derived from UBV magnitudes. These temperature trends are clearly contributing to the dispersion of [O/H] results and it is to be hoped that fully lineblanketed models and line-formation calculations will resolve this issue and lead to a better zero-point for the composition of local B-type stars. It is certainly worth noting, however, that the long-standing discrepancy between oxygen abundances for the B-type stars and the Sun is largely removed by improvements in the Solar estimates, the latest three-dimensional results agreeing very nicely with those for the massive stars!

Was this article helpful?

0 0

Post a comment