Context. The identification and characterisation of populations of young massive stars in (giant) H ii regions provides important constraints on i) the formation process of massive stars and their early feedback on the environment, and ii) the initial conditions for population synthesis models predicting the evolution of ensembles of stars. Aims. We identify and characterise the stellar populations of the following young giant H ii regions: M8, G333.6-0.2, and NGC 6357. Methods. We have acquired H-and K-band spectra of around 200 stars using The K-band Multi Object Spectrograph (KMOS) on the ESO Very Large Telescope. The targets for M8 and NGC 6357 were selected from the Massive Young Star-Forming Complex Study in Infrared and X-ray (MYStIX), which combines X-ray observations with near-infrared and mid-infrared data. For G333.6-0.2, the sample selection is based on the near-infrared colours combined with X-ray data. We introduce an automatic spectral classification method in order to obtain temperatures and luminosities for the observed stars. We analysed the stellar populations using their photometric, astrometric, and spectroscopic properties and compared the position of the stars in the Hertzprung-Russell diagram with stellar evolution models to constrain their ages and mass ranges. Results. We confirm the presence of candidate ionising sources in the three regions and report new ones, including the first spectroscopically identified O stars in G333.6-0.2. In M8 and NGC 6357, two populations are identified: (i) OB main-sequence stars (M > 5 M ) and (ii) pre-main sequence stars (M ≈ 0.5 − 5 M ). The ages of the clusters are ∼1-3 Myr, < 3 Myr, and ∼0.5-3 Myr for M8, G333.6-0.2, and NGC 6357, respectively. We show that MYStIX selected targets have > 90% probability of being members of the H ii region, whereas a selection based on near infrared (NIR) colours leads to a membership probability of only ∼70%. Arias et al. (2007) performed intermediate resolution spectroscopy of PMS stars in M8. They identified 27 classical T Tauri stars, seven weak-lined T Tauri stars and three PMS emission objects with spectral type G. They identified the stars to be younger than 3 Myr and between 0.8 and 2.5 M . Prisinzano et al. (2007, 2019) identified 237 PMS stars as members including 53 binaries. Kumar & Anandarao (2010) identified 64 class 0/I and 168 class II objects using Spitzer near-infrared photometry. Article number, page 2 of 39 M.C. Ramírez-Tannus et al.: The young stellar content of M8, G333.6-0.2, and NGC 6357.

The majority of massive stars (> 8 M⊙) in OB associations are found in close binary systems. Nonetheless, the formation mechanism of these close massive binaries is not understood yet. Using literature data, we measured the radial-velocity dispersion (σ1D) as a proxy for the close binary fraction in ten OB associations in the Galaxy and the Large Magellanic Cloud, spanning an age range from 1 to 6 Myr. We find a positive trend of this dispersion with the cluster’s age, which is consistent with binary hardening. Assuming a universal binary fraction of fbin = 0.7, we converted the σ1D behavior to an evolution of the minimum orbital period Pcutoff from ∼9.5 years at 1 Myr to ∼1.4 days for the oldest clusters in our sample at ∼6 Myr. Our results suggest that binaries are formed at larger separations, and they harden in around 1 to 2 Myr to produce the period distribution observed in few million year-old OB binaries. Such an inward migration may either be driven by an interaction with a remnant accretion disk or with other young stellar objects present in the system. Our findings constitute the first empirical evidence in favor of migration as a scenario for the formation of massive close binaries.

Context. Recently much progress has been made in probing the embedded stages of massive star formation, pointing to formation scenarios that are reminiscent of a scaled-up version of low-mass star formation. However, the latest stages of massive-star formation have rarely been observed, as young massive stars are assumed to reveal their photospheres only when they are fully formed. Aims. Using first and second overtone CO bandhead emission and near- to mid-infrared photometry, we aim to characterize the remnant formation disks around five unique pre-main-sequence (PMS) stars with masses 6–12 M⊙ that have constrained stellar parameters thanks to their detectable photospheres. We seek to understand this emission and the disks from which it originates in the context of the evolutionary stage of the studied sources. Methods. We used an analytic disk model, and adopted local thermodynamical equilibrium, to fit the CO bandhead and the dust emission, assumed to originate in different disk regions. For the first time, we modeled the second overtone emission, which helped us to put tighter constraints on the density of the CO gas. Furthermore, we fit continuum normalized bandheads, using models for stellar and dust continuum, and show the importance of this in constraining the emission region. We also included 13CO in our models as an additional probe of the young nature of the studied objects. Results. We find that the CO emission originates in a narrow region close to the star (<1 AU) and under very similar disk conditions (temperatures and densities) for the different objects. This is consistent with previous modeling of this emission in a diverse range of young stellar objects and identifies CO emission as an indicator of the presence of a gaseous inner disk reaching close to the stellar surface. From constraining the location of the inner edge of the dust emission, we find that all but one of the objects have undisrupted inner dust disks. Conclusions. We discuss these results in the context of the positions of these PMS stars in the Hertzsprung-Russel diagram and the CO emission’s association with an early age and high accretion rates in (massive) young stellar objects. We conclude, considering their mass range and the fact that their photospheres are detected, that the M17 PMS stars are observed in a relatively early formation stage. They are therefore excellent candidates for longer wavelength studies to further constrain the end stages of massive star formation.