Astron. Astrophys. 331, 193-210 (1998)

5. Properties of the X-ray sources

About 3/4 of the ROSAT sources in the two clouds have optical counterparts, implying a relatively low average extinction reminiscent of the Chamaeleon I cloud (FCMG). The remaining sources without visible counterparts are candidate embedded young stellar objects, as seen in the Ophiuchi cloud core (CMFA). Unfortunately, little information is available on the proposed optical counterparts, and the deep IR surveys cover only small portions of the ROSAT field. For example, it is difficult to distinguish cloud members from field stars among the 10 12 counterparts. We therefore consider here the X-ray data alone, assuming that all sources are associated with the clouds, and checking this assumption a posteriori. We consider the individual X-ray properties, X-ray hardness ratios, and ratios.

5.1. Individual fluxes and luminosities

The brightest ROSAT PSPC sources have sufficient counts to fit X-ray spectra and derive individual absorption-corrected fluxes. Spectral analysis is based on the Raymond-Smith thermal plasma model where the source temperature , line-of-sight absorption expressed as the equivalent hydrogen column density , and the elemental composition are free parameters. We find the best-fit X-ray spectra have solar abundances, temperatures around 1 keV, and a wide range of column densities cm^-2. When the visual extinction for the stellar counterpart is available, it can be compared to the value of obtained from X-ray spectral fitting using the relation cm^-2 assuming (e.g., Ryter 1996). The sources generally have too few counts to allow an X-ray determination of ; however, for the stronger sources there is good agreement between the X-ray and optically derived column densities. For and 1 keV, the correspondence between count rate and X-ray flux is given by: 1 count/ksec = erg s^-1 cm^-2.

By adopting this extinction for all sources and the cloud distances given in Table 1, approximate values for the 0.4 - 2.4 keV X-ray luminosities, , can be derived. In Tables 2 and 3 we list the values and respective errors as concluded from the errors in the count rates and in the fit results above mentioned. The resulting luminosity distributions (Fig. 5) lie above the sensitivity limits of erg s^-1 (Monoceros) and erg s^-1 (Rosette), and below a maximum around erg s^-1. The fainter TTS with erg s^-1 which dominate the Chamaeleon I and Ophiuchi stellar populations (FCMG, CMFA) are undetectable at the greater distances of the Monoceros and Rosette clouds. While high-luminosity erg s^-1 sources are more numerous in the Rosette cloud than in the Monoceros cloud, the maximum luminosity in both fields is not significantly larger than for the nearby clouds. This immediately indicates that the brightest ROSAT sources cannot be embedded massive stars, which would be much more X-ray luminous.

Fig. 5a and b. Differential X-ray luminosity functions for the Monoceros and Rosette clouds. The lack of faint sources for the Rosette cloud compared with the Monoceros cloud is a distance effect. To minimize differential extinction effects, the approximate luminosities are evaluated in the spectral range [1.0 - 2.4] keV here and in the following figures.

While a quantitative comparison of the Monoceros and Rosette X-ray luminosity functions with those of nearby star forming regions would be very informative, we believe it cannot yet be reliably performed. The total stellar population of the more distant regions is unknown, so we do not know what fraction of the total population is represented by the observed distributions in Fig. 5. Thus, we cannot discriminate between intrinsic differences in luminosity function shapes (e.g., Rosette stars are more luminous than Monoceros stars) and differences in their normalizations (e.g., the Rosette population is larger than that in Monoceros).

5.2. X-ray hardness ratios

The study of X-ray sources within the Taurus molecular cloud by Neuhäuser et al. (1995b) using the ROSAT All-Sky Survey (RASS) shows that T Tauri stars can be distinguished from other sources using X-ray spectral hardness ratios and . This is an X-ray color-color diagram. In our study, the soft channels are removed and only is available, where are the count rates in the indicated energy ranges (in keV). Fig. 6 shows the curves as calculated from the XSPEC software package. They are almost independent of for cm^-2 (equivalent to ) but rise sharply for larger values. does not depend very much on for keV. For high values of , and for keV, . From Stark et al. (1992), the total line-of-sight log 21.3 and 21.8 in the direction of the Monoceros and Rosette clouds, respectively. The column densities to the front of the clouds must be less than these values. Taking keV for the temperature of TTS, expected ranges for can be obtained from Fig. 6: (Monoceros) and (Rosette). These results are in agreement with those found by Neuhäuser et al. (1995b, see their Fig. 4).

Fig. 6. Hardness ratio as a function of the equivalent hydrogen column density (bottom scale) or visual extinction (top scale), for different source temperatures .

Fig. 7 shows the observed distribution of for strong Monoceros and Rosette sources. For most sources, the ratio is low and the statistical uncertainties are comparable to the values of and may be biased because cannot exceed unity. We estimate that about half of the bright Monoceros and Rosette sources have high consistent cutoff spectra with cm^-2. The few sources with are soft and may be foreground stars unrelated to the cloud.

Fig. 7a and b. Histogram of the Monoceros and Rosette X-ray sources as function of their hardness ratio HR2

5.3. ratios

Pre-main sequence stars may be discriminated from main sequence stars by their high ratios: for late-type field stars; for Herbig Ae/Be stars (Zinnecker & Preibisch 1994); for T Tauri stars (FCMG); and higher for some protostars (Grosso et al. 1997). We roughly estimate the bolometric luminosities of stars in the present samples based on measured (Tables 2 and 3), assuming cloud membership, G spectral type and 1 magnitude absorption at R. The resulting relation between and or is displayed in Fig. 8. The straight line is the () correlation obtained for Cha I cloud stars (Lawson, Feigelson & Huenemoerder 1996). We have also added in Fig. 8 the X-ray data on Herbig Ae/Be stars (Zinnecker & Preibisch 1994), taking into account that, since these stars have earlier spectral types, they have a different color index and bolometric correction than G stars.

Fig. 8. X-ray luminosities plotted against absolute R magnitudes for optically visible Monoceros (triangles) and Rosette (squares) sources. Also shown for comparison are Cha I TTS (stars; upper limits are indicated by arrows) and Herbig Ae/Be stars (crosses). The full line shows the fit obtained for the Cha I sources, (erg s (Lawson et al. 1996).

We first notice that, with very few exceptions, the X-ray sources for the three clouds are all consistent (within uncertainties) with the Cha I (log ) correlation reported by FCMG. There is no clustering of sources two orders of magnitude below this line, as expected for field stars. If anything, many of the Monoceros and Rosette X-ray emitting stars tend to lie 0.5 to 1 dex above the correlation found for the Cha I sources; a few stars, especially in the Rosette cloud, even reach luminosities up to several times erg s^-1, corresponding to late B stars. We conclude that the Monoceros and Rosette ROSAT fields contain only a small number of unidentified field stars, and that the majority of sources must be young stars.

5.4. Summary: nature of the X-ray source population

The preceding sections have shown, by a variety of independent methods, that the populations of ROSAT sources in the Monoceros and Rosette clouds (except perhaps in the extended feature at the NW periphery of the Rosette field) are predominantly made up of young intermediate- and low-mass stars similar to Herbig Ae/Be stars and the more luminous T Tauri stars in the Cha I and Oph clouds. There is little contamination from field stars, and extragalactic contamination should be negligible. Without optical spectroscopy and photometry, we are unable to derive luminosities and masses from locations on the Hertzsprung-Russell diagram, to discriminate between "classical" and "weak" TTS, or to deduce the relative ages or underlying populations of the distant clouds.

© European Southern Observatory (ESO) 1998

Online publication: February 4, 1998
helpdesk.link@springer.de