2024A&A...690A.113B

Context. The outcome of the formation of massive stars is an important anchor point in the formation and evolution process of these stars. It provides insight into the physics of the assembly process, and sets the conditions for stellar evolution. For massive stars, the outcome of formation is rarely observed because the processes involved unfold deep down in highly extincted molecular clouds.
Aims. We characterize a population of highly reddened stars in the very young massive star-forming region M17. The group of 18 O4.5 to B9 stars constitutes one of the best samples of almost zero-age main-sequence (ZAMS) high- and intermediate-mass stars. Their properties allow us to identify the empirical location of the ZAMS of massive stars, and the rotation and mass-loss rate of stars close to or at the onset of core-hydrogen burning.
Methods. We performed quantitative spectroscopic modeling of a uniform set of over 100 spectral features in optical VLT/X-shooter spectra using the nonlocal thermal equilibrium stellar atmosphere code FASTWIND and a fitting approach based on a genetic algorithm, KIWI-GA. The spectral energy distributions of photometric observations were used to determine the line-of-sight extinction. From a comparison of their positions in the Hertzsprung-Russell diagram with MIST evolutionary tracks, we inferred the stellar masses and ages.
Results. We find an age of 0.4_–0.2^+0.6 Myr for our sample, however we also identify a strong relation between the age and the mass of the stars. All sources are highly reddened, with A_V ranging from 3.6 to 10.6 mag. The sample can be subdivided into two groups. Stars more massive than 10 M_⊙ have reached the ZAMS. Their projected ZAMS spin rate distribution extends to 0.3 of the critical velocity; their mass-loss rates agree with those of other main-sequence O and early-B stars. Stars with a mass in the range 3 < M/M_⊙ < 7 are still on the pre-main sequence (PMS), and some of them have circumstellar disks. Evolving their υ sin i to the ZAMS assuming angular momentum conservation yields values up to ∼0.6 υ_crit. For PMS stars without disks, we find tentative mass-loss rates up to 10^–8.5 M_⊙ yr^–1. The total-to-selective extinction R_V is higher for PMS stars with disks than for the remainder of the sample.
Conclusions. We constrain the empirical location of the ZAMS for massive (10 < M/M_⊙ < 50) stars and find it to agree with its location in MIST evolutionary tracks. The ZAMS rotation rates for intermediate-mass stars are twice as high as for massive stars, suggesting that the angular momentum gain processes differ between the two groups. The relation between the age and mass of the stars suggests a lag in the formation of more massive stars relative to lower mass stars. Taking the derived mass-loss rates at face value, stellar winds are already initiated in the PMS phase. The PMS-star winds are found to be substantially more powerful than indicated by predictions for line-driven outflows.