Daniel J. Lennon1'2 & Carrie Trundle3
1Isaac Newton Group of Telescopes, E-38700 Santa Cruz de La Palma, Tenerife, Spain 2Instituto de Astrofísica de Canarias, E-38200 La Laguna, Tenerife, Spain 3 The Queen's University of Belfast, Belfast BT71NN, Northern Ireland, UK
We discuss the metallicity of massive stars in the Solar neighbourhood, comparing new results with those for the Sun. We find that, despite there being small systematic differences between various NLTE determinations of [O/H] in hot stars, there is reasonable agreement among results from various studies of nearby stars, with a value of 8.60 ± 0.1 dex being implied. This is in good agreement with the latest Solar estimate based on three-dimensional models, and is in good agreement with recent estimates of the nebular oxygen abundance in Orion. We review the evidence for metal-rich massive stars in our own galaxy and in M31, concluding that there is little convincing evidence for supersolar [O/H] in massive stars in the Milky Way, while there is only limited evidence for mildly metal-rich regions in M31 with [O/H] relative to Solar of only +0.2. Discrepancies between stellar and nebular abundances at high metallicity can be traced to problems in calibrating the R23 index for H ii regions in the metal-rich regime.
Any discussion of metal-rich massive stars would be incomplete without first addressing the issue of the abundance scale in order to answer the questions 'What constitutes a metal-rich massive star?' We can then address the question 'Do we observe any metal-rich massive stars directly?'
Determining the abundance scale is important. For example, nearby massive stars are generally assumed to have 'Solar' composition, which is typically used in stellar-evolution calculations and formulations of mass-loss rate, and which in turn is scaled to other metallicities (Z). Historically, however, there have always been problems in resolving the chemical composition of nearby OB stars with that of the Sun, and to a large extent it was discrepancies such as this that drove much of the early development of NLTE model atmospheres for hot stars; see Mihalas (1972) for the case of magnesium. The expectation was that much of these apparent discrepancies was due to the neglect of NLTE and, in more recent times, to the neglect of line-blanketing and winds in the hotter and more luminous stars. So, while a differential analysis can result in precise relative abundances, even using LTE methods, NLTE techniques are required in order to set the zero-point of the abundance scale and enable us to compare with the Sun. In Section 2 we will review abundances for nearby B-type stars and compare these with the latest results from Solar three-dimensional models. It is interesting to note in this context that the accepted Solar composition has been changing as a result of the implementation of these new models and, as we will see, at least for oxygen, nearby massive stars and the Sun are converging to the same abundance.
Besides the Sun, H 11 regions provide another very important comparison for massive stars. In a sense this is a more relevant comparison since the chemical compositions of both massive stars and H11 regions should reflect that of the current interstellar medium. Massive stars and their associated ionized nebulae should therefore have the same composition (though perhaps modified slightly by condensation onto dust grains). Comparing these two classes of objects is important because H11 regions are widely used to infer the dependences of various observables on Z, such as the blue-to-red supergiant ratio, Wolf-Rayet populations and in general calibrate indirect methods of inferring Z. In Section 3 we will look at the comparison of OB stars in Orion with the nebular abundance. Then in Section 4 we will turn to the comparison of stellar with nebular abundances in those inner regions of the Milky Way and M31 thought to be metal-rich from studies of these galaxies' abundance gradients as derived from nebular analysis.
Finally, throughout his article when we refer to metallicity, or Z, we will in general mean the oxygen abundance (or [O/H]). It is difficult to derive accurate and precise iron abundances in massive stars, since most iron lines lie in the UV and, at Solar metallicity and above, the lines are saturated and insensitive to abundance, although it may well be possible to estimate metallicity from mass-loss rates. see Haser et al. (1998) for a detailed discussion on these points. Oxygen, on the other hand, is an excellent tracer of metallicity; there are many O ii lines in the optical spectra of stars of spectral types B2-O9 and therefore we will use [O/H] as a proxy for Z unless specified otherwise. This has the further advantage that oxygen abundances are readily obtainable from H ii regions, indeed often this is the only abundance readily derived from faint and distant H ii regions.
Table 6.1. Mean oxygen abundances ([O/H]) for B2-O9 stars within 500pc of the Sun, the sample size (N) is also given. For comparison two Solar values are shown; the old value (Anders & Grevesse 1989), which is weighted towards analysis of the forbidden [O i] line using one-dimensional models, and the new results using a three-dimensional model (Asplund et al. 2004).
Solar, one-dimensional 8.93 ± 0.04
Solar, three-dimensional 8.66 ± 0.05
Was this article helpful?