http://arxiv.org/abs/1312.7234
The reliability of C, N, and O abundances determined from the C II, N II, and O II lines in the main sequence (MS) early B- and late O-type stars is examined. This analysis is based on a fact of primary importance for this problem: the C II, N II, and O II ions in atmospheres of these stars are photoionized by radiation in the far UV. Observations show that the actual flux in this range for hot stars can be considerably higher than theoretical levels; thus, overionization of the C II, N II, and O II ions beyond the ionization calculated with the standard model atmospheres occurs in the actual atmospheres of these stars. Underestimates in the calculated ionization lead to reduced values of the C, N, and O abundances. This is confirmed by the observed dependence on the effective temperature Teff: the C, N and O abundances tend to be lower with increasing Teff. It is shown that overionization becomes significant for sufficiently hot stars, namely when Teff>18500 K for the C II lines and when Teff>26000 K for the N II and O II lines. The systematic difference in the C, N and O abundances between these relatively hot stars and the cooler B-type MS stars is about 0.2 dex. The relatively cool B-type MS stars (Teff<18100 K for C II and Teff<25000 K for N II and O II) yield the undistorted C, N and O abundances, and their average values agree very well with modern estimates for the Sun. It is noted that when an alternative method based on examining the ionization balance (e.g., He I-II, C II-III, Ne I-II, Si III-IV) is used to determine the parameters Teff and log g, the overionization problem essentially vanishes; however, this method leads to a systematic enhancement in Teff and log g. The overionization problem can be solved by shifting to the improved model atmospheres for early B- and late O-type stars that takes their sphericity into account and, possibly, stellar winds and magnetic fields.
Mon, 30 Dec 13
14/26