Massive stars in the field and associations of the magellanic clouds: The upper mass limit, the initial mass function, and a critical test of main-sequence stellar evolutionary theory

Abstract

We investigate the massive star population of the Magellanic Clouds with an emphasis on the field population, which we define as stars located further from any OB association than massive stars are likely to travel during their short lifetimes. The field stars must have been born as part of more modest star-forming events than those that have populated the large OB associations found throughout the Clouds. We use new and existing data to answer the following questions: Does the field produce stars as massive as those found in associations? Is the initial mass function (IMF) of these field massive stars the same as those of large OB complexes? How well do the Geneva low-metallicity evolutionary models reproduce what is seen in the field population, with its mixed ages? To address these issues we begin by updating existing catalogs of LMC and SMC members with our own new spectral types and derive H-R diagrams (HRDs) of 1584 LMC and 512 SMC stars. We use new photometry and spectroscopy of selected regions in order to determine the incompleteness corrections of the catalogs as a function of mass and find that we can reliably correct the number of stars in our HRDs down to 25 M_sun_. Using these data, we derive distance moduli for the Clouds via spectroscopic parallax, finding values of 18.4 +/- 0.1 and 19.1 +/- 0.3 for the LMC and SMC. The average reddening of the field stars is small: E(B - V) = 0.13 (LMC) and 0.09 (SMC), with little spread. We find that the field does produce stars as massive as any found in associations, with stars as massive as 85 M_sun_ present in the HRD even when safeguards against the inclusion of runaway stars are included. However, such massive stars are much less likely to be produced in the field (relative to lower mass stars) than in large OB complexes: the slope of the IMF of the field stars is very steep, {GAMMA} = -4.1 +/- 0.2 (LMC) and {GAMMA} = - 3.7 +/- 0.5 (SMC). These may be compared with {GAMMA} = - 1.3 +/- 0.3, which we rederive for the Magellanic Cloud associations. (We compare our association IMFs with the somewhat different results recently derived by Hill et al. and demonstrate that the latter suffer from systematic effects due to the lack of spectroscopy.) Our reanalysis of the Garmany et al. data reveals that the Galactic field population has a similarly steep slope, with {GAMMA} = - 3.4 +/- 1.3, compared to {GAMMA} = - 1.5 +/- 0.2 for the entire Galactic sample. We do not see any difference in the IMFs of associations in the Milky Way, LMC, and SMC. We find that the low metallicity evolutionary tracks and isochrones do an excellent job of reproducing the distribution of stars in the HRD at higher masses, and in particular match the width of the main-sequence well. There may or may not be an absence of massive stars with ages less than 2 Myr in the Magellanic Clouds, as others have found for Galactic stars; our reddening data renders unlikely the suggestion that such an absence (if real) would be due to the length of time it takes for a massive star to emerge. There is an increasing discrepancy between the theoretical ZAMS and the blue edge of the main- sequence at lower luminosities; this may reflect a metallicity dependence for the intrinsic colors of stars of early B and later beyond that predicted by model atmospheres, or it may be that the low metallicity ZAMS is misplaced to higher temperatures. Finally, we use the relative number of field main-sequence and Wolf-Rayet stars to provide a selection-free determination of what mass progenitors become WR stars in the Magellanic Clouds. Our data suggest that stars with initial masses > 30 M_sun_ evolve to a WR phase in the LMC; while the statistics are considerably less certain for the SMC, they are consistent with this limit being modestly higher there, possibly 50 M_sun_, in qualitative agreement with modem evolutionary calculations.