A Lifetime of Influence: Abstracts, Talks, and Posters
 

Discordant Estimates of Mass-Loss Rates for O-Type Stars

Alex Fullerton
STScI/HIA

In their influential paper "Mass Loss from O-Type Stars," Conti & Garmany (1980, ApJ, 238, 190) noted that while "unequivocal evidence of mass loss can come from the violet-shifted, strong resonance lines in the UV spectrum", for determinations of the mass-loss rate "the most crucial parameter, and the hardest to determine with these data, is the ionization fraction."

These insights from the early days of IUE succinctly foreshadow the limited role that UV resonance lines have come to play in mass-loss studies. On the one hand, resonance lines are certainly the most sensitive indicators of stellar winds. But on the other hand, quantitative determinations of mass-loss rates from them continue to be plagued by uncertainties in ionization fractions. This is particularly true for the wind lines accessible to IUE and HST, which are either due to (a) dominant ions of abundant elements, which invariably produce saturated lines; or (b) trace ions whose abundances depend sensitively on the action of additional processes (e.g., line blocking; X-rays). Consequently, a strong preference has developed for mass-loss rates derived from observations of emission in Halpha or the free-free continuum at radio wavelengths.

The rich suite of far-ultraviolet resonance lines accessible to FUSE and its predecessors alters this situation by providing additional diagnostic leverage for mass-loss determinations. The P V resonance doublet is particularly useful, because (a) for a subset of O temperature classes, it traces the dominant ion in the wind, which implies that its ion fraction is known a priori to within a factor of ~2; and (b) even when its parent ion is dominant, the P Cygni absorption troughs are not saturated because P is such a rare element. However, even though the P V doublet should provide reliable mass-loss rates for at least some O stars, surveys of Galactic and Magellanic O stars show that the values derived from P V are systematically smaller than those determined from Halpha or radio measurements, typically by an order of magnitude or more. These disparate measurements can be reconciled if the winds of O-type stars are assumed to be significantly clumped on small spatial scales, since "density squared" emission processes over-estimate the amount of material if the outflow is inhomogeneous. Insofar as the preferred mass-loss rates are currently based "density squared" processes, it follows that they are too large.

I will review current understanding of the discordant mass-loss estimators of O-type stars, and will outline recent progress toward resolving the discrepancy. Although definitive estimates of the reduction in mass-loss rates due to clumping are not yet available, concordance appears to be within reach. Open issues concerning the structure of line-driven stellar winds will also be discussed.