3 Selection

3.1 Extinction map of the Cha II cloud

We construct an extinction map of the Cha II using the star count method developed by Cambrésy (1999). This method is based on the comparison of local stellar densities in the absorbed region and a nearby reference area assumed to be free of obscuration. DENIS provides stars detected in 2 infrared bands and 1 optical band. We first need to determine which band is the best one to apply this method. The $K_{\rm s}$ band allows us to probe the dense cores of the cloud but the density contrast is too small to build an extinction map. We therefore use the I and J bands to construct the extinction map. By requiring that the stars be detected in both bands, we can eliminate spurious sources. But the simultaneous use of I and J bands introduces a bias which begins at $A_V = [(I-J)_{\rm limit} - (I-J)_{\rm average}]$/[<A_I/A_V>- <A_J/A_V>]. Figure 2 shows that the average color of the stars $(I-J)_{\rm average}=1$. Using the values of <A_I/A_V> and <A_J/A_V> from Cardelli et al. (1989), we find that the bias begins at A_V = (2-1)/(0.479-0.282)=5. The bias can reach 2 mag in the densest cores. As a result, fewer stars will be selected with a red color criterion. We use the $\sim$70000 sources detected in I and J bands to construct the extinction map. We apply a wavelet transform algorithm to filter out the noise due to the Poisson fluctuations in the star counts (Cambrésy 1999). We take into account the decrease of stellar density with latitude to calibrate the extinction as a function of galactic latitude. The slope of A_V(b) is equal to 0 as suggested by the morphological similarity with the ¹³CO map of Mizuno et al. (1999). The absolute calibration of the extinction is obtained by assuming the extinction is 0 outside of the cloud. We use the extinction curve of Cardelli et al. (1989) to convert A_J into A_V ( $A_J= 0.282 \times A_V$).

An alternative method to determine the extinction of individual stars consists of assuming an intrinsic color of stars and comparing this with the observed colors. However this method requires an assumption about the spectral types of the stars, for which we have no a priori information. The advantage of the star count method is that it provides an independent determination, at the expense of spatial resolution.

Figure 1 shows the resulting extinction map of the Cha II with a spatial resolution of 2 $\hbox{$^\prime$ }$. Note that Fig. 1 closely resembles the ¹³CO map of Mizuno et al. (1999). The extinction values A_V range from 0 to 12. The statistical magnitude uncertainty due to the number of stars counted is approximately given by $1.3/\sqrt{(n+1)}$ (Rossano 1980), where n is the number of stars counted per sampling element. Here, we have used n=20 and the statistical uncertainty is $\sim$0.30 mag in A_J ($\sim$1.0 in A_V). We use this map to deredden all stars detected by DENIS in 3 bands $IJK_{\rm s}$ in this area.

Figure 1: Spatial distribution of low-mass TTS and/or young brown dwarf candidates (triangle signs) selected with $I-J \ge 2$ in the Cha II cloud. Six candidates cross-identified with ROSAT X-ray sources are presented. The plus signs represent 15 bright candidates ( $K_{\rm s} \sim 8$). Contours indicate the extinction values beginning at AV=1 and spaced by 1 as derived from DENIS I and J counts. Coordinates are J2000.

Figure 2: $I-J/J-K_{\rm s}$ color-color dereddened diagram for new DENIS selected low-mass TTS and/or young brown dwarf candidates (triangle signs) in Cha II cloud. The dots represent 20 000 stars detected in DENIS $IJK_{\rm s}$ bands in an area of $2.42\hbox {$^\circ $ }\times 3.50\hbox {$^\circ $ }$ of the Cha II. The open dot signs represent known TTS in the Cha II. Eleven confirmed young brown dwarfs in the Cha I cloud are also shown (Comerón et al. 2000). The plus signs are 15 bright candidates ( $K_{\rm s} \sim 8$). The main sequence, the giant branch and the extinction vector are also plotted. The representative error bar is shown in the upper right corner.

3.2 Selection criterion

As shown by Delfosse (1997), the I-J versus $J-K_{\rm s}$ diagram is a powerful tool to separate low-mass TTS and young brown dwarfs from main sequence stars. Brown dwarfs and low-mass TTS have no hydrogen fusion when they are young. Thus, the nature of these objects is the same, and only their mass and/or temperature are different. This explains why they are both characterised by a deficit in I band because of the molecular absorption bands. Their youth is expressed by H$\alpha$ emission and Li  I absorption lines. But they are also characterised by features such as TiO and VO bands, and CaH, which allow the evaluation of the temperature and surface gravity. Kirkpatrick et al. (1991) have shown that these bands are very sensitive to the temperature of the environment in the M dwarfs. Because these features are located in the I-band, these objects have a strong I band deficit, while the J band flux is essentially photospheric in all these objects. Therefore, I-J is a good estimator of their effective temperature. Thanks to the association of the optical I band to the infrared J and $K_{\rm s}$ bands in DENIS, we have derived a $I-J/J-K_{\rm s}$ diagram (see Fig. 2) for $\sim$20 000 sources detected in three bands $IJK_{\rm s}$. We only select sources after dereddening with $I-J \ge 2$ and spatially distributed in groups around the cloud cores. This provides a sample of 98 sources. After visual inspection of the DENIS images, we exclude 28 objects because they are located near the bleed-out trails of nearby saturated stars. We note that this selection criterion targets only young brown dwarfs and low-mass TTS. Massive TTS are characterised by their excess in $K_{\rm s}$, thus in $J-K_{\rm s}$ and $H-K_{\rm s}$. Brown dwarfs and low-mass TTS have important I-J but normal $J-K_{\rm s}$.

Figure 3: J vs. I-J magnitude-color diagram for the Cha II dark cloud. Selected low-mass TTS and/or young brown dwarf candidates are indicated with triangles. We show the 1 Myr, 10 Myr and 100 Myr isochrones (dashed lines) from the Baraffe et al. (1998) model. Open circles denote known TTS of the Cha II. The star symbols represent brown dwarfs detected and spectroscopically confirmed by Comerón et al. (2000) and the plus signs, 15 bright candidates also selected with $I-J \ge 2$. The representative error bar is given.

Table 1 lists the remaining 70 low-mass young star candidates with their intrinsic photometric uncertainties, i.e. these do not include the uncertainties in the derived dereddening from the extinction map. Note that due to the bias (Sect. 3.1) introduced by I and J selection in the extinction map, fewer stars are selected in regions of high extinction. We identify 4 known TTS and 7 sources detected by IRAS within a 1 $\hbox{$^\prime$ }$ radius in this sample. Four of our candidates are also detected by ISO (P. Persi, private communication). Figure 2 shows these candidates, and 11 spectroscopically confirmed brown dwarfs detected in the Cha I cloud (Comerón et al. 2000). We note that most of our candidates are located in the same region of the diagram as these known brown dwarfs.

3.3 Cross-identification with ROSAT sources

Weak line TTS are known as objects without a disk. Because of their relatively weak emission in the infrared (ascribed to the lack of dense surrounding matter), they can be more easily detected by their photospheric or coronal emission in X-ray surveys. Recently, Alcalá et al. (2000), using ROSAT PSPC observations, have detected 40 X-ray sources in the Cha II cloud, of which only 14 have been identified with previously known young stellar objects (IRAS sources, TTS). We cross-identified these 40 sources with DENIS $IJK_{\rm s}$ data, and identified these 14 known objects, plus 6 additional new WTTS candidates. We have also checked that these sources are located in the cloud (see Fig. 1).