Michele D. Thornley

Massive Star Formation and Evolution in Starburst Galaxies: Mid-infrared Spectroscopy with the ISO Short Wavelength Spectrometer*

We present new Infrared Space Observatory Short Wavelength Spectrometer data for a sample of 27 starburst galaxies, and with these data we examine the issues of formation and evolution of the most massive stars in starburst galaxies. Using starburst models which incorporate time evolution, new stellar atmosphere models for massive stars, and a starburst model geometry derived from observations of the prototypical starburst M82, we model the integrated mid-infrared line ratio (Ne III)(15.6 km)/(Ne II)(12.8 km). This line ratio is sensitive to the hardness of the stellar energy distribution and therefore to the most massive stars present. We conclude from our models, with consideration of recent determinations of the stellar census in local, high-mass star-forming regions, that the (Ne III)/(Ne II) ratios we measure are consistent with the formation of massive (D50¨100 stars in most starbursts. In this framework, M _ ) the low nebular excitation inferred from the measured line ratios can be attributed to aging eUects. By including estimates of the ratio of infrared-to-Lyman continuum luminosity for the galaxies in our sample, we further —nd that most starbursts are relatively short-lived (106¨107 yr), only a few O star lifetimes. We discuss a possible cause of such short events: the eUectiveness of stellar winds and super- novae in destroying the starburst environment. Subject headings: galaxies: starburstinfrared: galaxiesstars: atmospheresstars: evolution ¨ stars: formationtechniques: spectroscopic

Star Formation Histories of Starbursts

We present results of a mid-infrared ISO spectroscopic survey of 27 starburst galaxies, and of near-infrared integral field spectroscopy and mid-infrared ISO spectroscopy of the starburst galaxy M 82. Together with the application of starburst models, the data are consistent with the formation of massive stars (≳50 – 100 M⊙) in starburst environments, and support short decay timescales of starburst episodes (~ 106 – 107 yr) indicating important negative feedback effects of starburst activity.