2000A&A...359..269C

We present observations of a sample of 13 very low mass stars and brown dwarfs in the central region of the Chamaeleon I star forming cloud. The observations include slitless spectroscopy around Hα to identify new members, low resolution long-slit visible and near-infrared spectroscopy, deep ROSAT PSPC X-ray observations, and ISOCAM mid-infrared observations. Our sample adds seven new objects to those discussed by Comeron, Rieke, and Neuhaeuser (1999A&A...343..477C) and extends the range of spectral types up to M8. We study different narrow-band indices as a tool for detecting and classifying very late-type young stellar objects. As to K-band spectra, we find that the visible features are not appropriate to yield a spectral classification more accurate than a few subclasses at best beyond M6. None of our sources displays K-band excess emission, but four have excess at 6.7µm suggesting that, although circumstellar disks are common around young very low mass stars, their inner regions are in general not hot enough to radiate significantly in the K band. Mid-infrared emission loosely correlates with Hα emission: sources without mid-IR excesses are always weak Hα emitters, while mid-IR excess sources have a broad range of Hα equivalent widths. X-ray emission is detected for 7 objects with spectral type M6 or later, including one bona-fide brown dwarf and three objects near the border separating stars and brown dwarfs. X-ray to bolometric luminosity ratios are typical of low mass, fully convective stars. The non-detection of X-ray emission at comparable levels from more evolved brown dwarfs suggests that X-ray activity may be restricted to early stages of brown dwarf evolution. We discuss in detail the temperatures and luminosities of our objects based on their magnitudes and spectra, and use the derived values to estimate masses and ages according to two different sets of pre-main sequence evolutionary tracks. Both sets of models are in good agreement concerning the mass derived for our objects, showing that four of them are bona-fide brown dwarfs, six are transition objects, and three are low mass stars. Derived ages differ significantly depending on the adopted models, especially at the lowest masses. This is mainly due to the objects lying on opposite sides of the deuterium-burning main sequence depending on whether one or another set is used. Using Baraffe et al. (1998A&A...337..403B) models for the dating of each object in the area of our survey with mass below 1 M_☉, we find that most have ages near 2x10⁶ years, with a small spread around that value. However, a few objects appear to have ages near 2x10⁷ years, suggesting that most, but not all, star formation in that region of Chamaeleon I may have happened almost simultaneously in a recent burst. Comparing predictions on members of the star forming region based on K-band star counts with the number of members actually identified through Hα emission suggests that sensitive Hα surveys are very efficient in producing a complete or nearly complete magnitude-limited census of young stellar objects in Chamaeleon I. Under the assumption that our sample is complete, we derive a mass function of Chamaeleon I between 0.03 and 1M_☉ which can be approximated by a nearly flat powerlaw in logarithmic mass units, in agreement with results for other young aggregates.