5 Summary and conclusions

In this paper we have discussed the results of a spectroscopic survey in the H and K bands of stars in the direction of Cygnus OB2, within a radius of one degree from its nucleus. The sample was selected on the basis of their JHK photometry as published in the Second Incremental Release of the 2MASS Point Source Catalog. Our selection is based on the fact that the color-color diagram of Cygnus OB2 shows two clearly distinct sequences, both parallel to the reddening vector: the lower one (i.e., bluer J-H at a given H-K) corresponds to early-type, moderately reddened members of the association, while the upper and most populated one corresponds to background, reddened giants. Some objects appear on or near the lower reddening sequence and large values of J-H and H-K and may be obscured or embedded members of Cygnus OB2.

We obtained spectra for stars brighter than K = 10 near the lower sequence and with a position in the K, (H-K) diagram consistent with them being O-type stars. The low resolution and moderate signal-to-noise ratio of our spectra does not allow an accurate spectral classification, but suffices to distinguish between the featureless spectra of early-type stars and those of cool stars, displaying a different continuum shape and prominent molecular absorption bands due to CO and H₂O. Intermediate-type stars should be too faint at the distance of Cygnus OB2 to be included in our target list.

We have identified 77 stars that we tentatively classify as early-type based on the shape of the continuum and the lack of visible features in our spectra. Only 31 have been previously classified based on visible spectra, 24 of them being O-type and 7 early B-type. Assuming that the same O- to B- star ratio holds for all the 77 stars, we thus estimate that there are approximately 60 O stars in this sample.

We also identify 20 stars with Br$\gamma$ emission in their K-band spectra, and in many cases other lines (mainly of H and He) in emission as well. These features are typical of the spectra of evolved massive stars undergoing intense mass loss in stellar winds. Of these, 16 are identified as emission-line stars for the first time in this study. Most of these stars are highly reddened, with 8 of them having H-K > 1 and reaching up to 3.1. Most of these stars, and all those having H-K > 1, appear to the right of the reddening vector, indicating the existence of circumstellar emission at long wavelengths. Based on the JHKcolors and fluxes, three of these stars, unidentified so far, seem to be among the most luminous members of Cygnus OB2. Three additional bright red stars showing CO overtone and infrared excess emission (one also showing H₂ S(1) emission at 2.12 $\mu$m) are also identified. Finally, we consider the possible existence of red supergiant members of Cygnus OB2 in our sample. Although we do find several stars with infrared spectra suggestive of them being supergiants, they are all too faint to be members of Cygnus OB2. This does not rule out the possibility that Cygnus OB2 does contain late-type supergiants, as our target selection is biased against these stars.

The location of the Cygnus OB2 members identified in this study, both the new ones and the previously known ones, confirm the shape of Cygnus OB2 as being roughly circular with an extension towards the Southwest, as determined by Knödlseder (2000) using star counts. We point out a possible difference in the distribution of normal early-type stars and that of emission-line stars, the latter having a more regular distribution around the core of the association, although the statistical significance of this result is limited by the small number of objects on which it is based.

Taking into account our results, those of other studies that have provided spectral classifications in the visible, and approximate corrections for incompleteness, we estimate that Cygnus OB2 contains between 90 and 100 O-type stars or stars having evolved from a O-type progenitor, in very good agreement with recent estimates by Knödlseder (2000). Of these, approximately 25% have emission lines in their infrared spectra.

Since its discovery over 40 years ago Cygnus OB2 has been recognized as an outstanding association, with new studies only confirming or enhancing its richness. While photometric and spectroscopic surveys have been hampered by the extinction hiding from view a significant part of the association, recent efforts based on the 2MASS near-infrared survey are providing an accurate estimate of the massive star contents of Cygnus OB2. The good agreement found between the independent approaches followed by Knödlseder (2000) and the present work indicates that the census of massive stars in Cygnus OB2 is nearing completion now, at least as far as the region within one degree from the center of the association is concerned. Such a complete census of by far the richest OB association in the solar neighbourhood, and probably one of the richest in our Galaxy, presents us with a unique opportunity to study in close-up detail its entire high-mass end and the wide variety of phenomena associated to the activity of very massive stars. High resolution infrared spectroscopy at high signal-to-noise ratio of all the objects listed in this study, as well as visible spectroscopy of the new, less reddened members, should provide a most valuable observational resource for understanding the formation, evolution, and impact on their environments of large OB associations in galactic disks, as well as the complex spectral variety resulting from the late evolutionary stages of the most massive stars.

Acknowledgements

We warmly thank the organizers of the first NEON Summer School, especially Dr. Michel Dennefeld, for their highly successful effort in making it a very profitful scientific and learning experience. The rest of participants and faculty of the School are also thanked for having made possible the unique human atmosphere during the two weeks of the School. The NEON Summer Schools are funded by the European Union. We also acknowledge the generous allocation of time at the Calar Alto 1.23 m telescope in the summer of 2001, which allowed us to follow up on this project and obtain the considerable number of spectra discussed in this paper. Constructive comments by an anonymous referee are thanked as well. Finally, it is a great pleasure to thank all the staff of the Calar Alto Observatory, and especially Mr. Santos Pedraz, for their support in our observing runs. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France, as well as of data products from the Two Micron All Sky Survey, which is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation.