Thanks to their usefulness in various fields of astrophysics (e.g. mixing processes in stars, chemical evolution of galaxies), the last few years have witnessed a large increase in the amount of abundance data for early-type stars. Two intriguing results emerging since the last reviews on this topic (Herrero 2003;Herrero & Lennon 2004) will be discussed: (a) nearby OB stars exhibit metal abundances generally lower than the solar/meteoritic estimates; (b) evolutionary models of single objects including rotation are largely unsuccessful in explaining the CNO properties of stars in the Galaxy and in the Magellanic clouds.Session: Atmospheres, mass loss and stellar winds
-IntroductionThis review about the chemical properties of massive stars will focus on two issues that are relevant in the context of this meeting. First, one may wonder how the chemical abundances of nearby OB stars compare to the solar values. This piece of information is, for instance, needed to properly model B-type pulsators and to draw correct inferences about their internal structure. Second, the abundances of several elements are powerful probes of mixing phenomena and, as such, can be used to improve our theoretical understanding of these processes and ultimately better model the evolution of massive stars across the HR diagram. Fascinating physical phenomena such as mass-transfer processes can affect the abundances of stars in binaries, but we refrain from discussing these systems here (see Langer, these proceedings).
-Getting the abundancesA prerequisite to obtain reliable abundances is to adopt accurate atmospheric parameters (a favourable case is offered by detached eclipsing binaries; e.g. Pavlovski & Southworth 2008). The effective temperature can be derived from photometric indices or, preferably, through ionisation balance of some metals (usually Si). The surface gravity is derived by fitting the collisionally-broadened wings of the Balmer lines, while the microturbulent velocity is inferred by requiring the abundances of a given ion to be independent of the line strength. Model atmospheres (either LTE or NLTE) with adequate line-blanketing are required. Departures from LTE are significant in hot stars and a full NLTE treatment for the line formation is also needed. Plane-parallel (e.g. TLUSTY, DETAIL/SURFACE) or so-called unified codes (e.g. CMFGEN, FASTWIND) can be used depending on the strength of the stellar wind. Of course, more sophisticated analysis techniques (NLTE model atmosphere, spherical extension) are much more demanding in terms of computer resources and should be used with discernment. A hybrid approach involving hydrostatic, LTE model atmospheres coupled with an NLTE