3. The age of the Tuc association (Comparison with other star-forming regions)
In this section the XLDF of the Tucanae candidates is studied. The X-ray emission declines with stellar age. Therefore, a comparison to XLDFs of other nearby regions of star formation should provide insight into the evolutionary stage of this association. In the following we describe the comparison samples.
3.1. TW Hydrae
Until the discovery of the Tucanae association, the TW Hydrae association was the only recognized nearby young stellar association far from molecular clouds. Strong X-ray emission is considered as (one of many) indicators for membership to the association since it was first studied in X-rays by Kastner et al. (1997). In subsequent studies which have added more stars to the group X-ray luminosities for the new members have been presented (see e.g. Jensen et al. 1998, Hoff et al. 1998, Webb et al. 1999, Sterzik et al. 1999, Hoff 2000). However, a systematic study of the X-ray properties of the whole sample has not been performed, and no XLDFs have been examined so far. In view of the similarity of the Tucanae and TW Hydrae regions, particularly the close and similar distance, it is intimating to compare their X-ray characteristics.
Fourteen TTS systems are known so far in the TW Hya region (see Rucinski & Krautter 1983, de La Reza et al. 1989, Gregorio-Hetem et al. 1992, Kastner et al. 1997, Jensen et al. 1998, Webb et al. 1999, Sterzik et al. 1999, Hoff 2000), all listed in Table 4. The multiplicities of these objects are given in Webb et al. (1999), Sterzik et al. (1999), and Hoff (2000). The possible spectroscopic binaries listed by Webb et al. (1999) have been confirmed as such (Torres et al., in prep.). The recently detected faint object next to TWA-7 (Neuhäuser et al. 2000a) has most recently been found to be a background K dwarf according to an H-band spectrum taken with ISAAC at the VLT (Neuhäuser et al. 2000b) so that we regard TWA-7 as single.
Table 4. RASS X-ray data of stars in the TW Hydrae association. X-ray data are derived from the RASS Bright Source Catalog with the exception of RX J1121.1-3845 (GSC 7739 2180), which is in a region of low exposure and therefore has no entry in the BSC. For this star we have extracted and analyzed the RASS raw data. The distances of the stars are computed from the Hipparcos parallax when available, otherwise we adopt a value of 55 pc, the mean of the Hipparcos distances measured for four TW Hya members. Multiples are listed only once. The meaning of the columns is the same as in Table 2.
We have cross-correlated the list of TW Hydrae stars with the RASS BSC. All but one star (GSC 7739 2180) can be identified with an X-ray source at less than from the optical position. In order to obtain the X-ray properties of GSC 7739 2180 which had no counterpart in the BSC we have checked the raw RASS data, and found that the star is located in a region of low exposure. We have detected an X-ray source with 10.7 cts at the optical position (RX J1121.1-3845). Thus, all TW Hydrae stars are X-ray emitters. Their X-ray characteristics are listed in Table 4. We give the designation of the star (Column 1), the X-ray position (Columns 2 and 3), the distance between optical and X-ray position (Column 4), the exposure time (Column 5), the broad band count rate (Column 6), the multiplicity (Column 7), X-ray luminosity (Column 8) and hardness ratios (Columns 9 and 10). As for the Tucanae stars, the luminosity of multiples has been divided by the number of components. Our X-ray data agree well with previously published ROSAT data for TWA stars 1.
Taurus-Auriga is one of the nearest ( pc; Elias 1978, Wichmann et al. 1998) and best studied regions of star formation. The region is particularly rich in late-type PMS stars. In an analysis of RASS data Neuhäuser et al. (1995) found that the subclass of weak-line TTS (i.e. TTS with weak H emission lines) in Taurus-Auriga are X-ray brighter than classical TTS (which are defined by equivalent widths of H). Because the Tucanae members appear to be somewhat evolved, they are probably all weak H emitters (naked weak-line post-TTS on radiative tracks near the ZAMS). Therefore, for our comparison we select only the weak-line TTS from Taurus-Auriga. We also note, that the classical TTS in Taurus-Auriga appear to be less X-ray bright than the wTTS although being younger (Neuhäuser et al. 1995, Stelzer et al. in prep). This may be due to magnetic star-disk coupling during the cTTS phase which prevents spin-up, and restricts the dynamo efficiency. The X-ray luminosity, therefore, shows a peak at the stage of the wTTS, and the decrease in X-ray luminosity with age sets in only after the disk is dissipated.
We have re-computed the XLDF for the RASS data of Taurus-Auriga weak-line TTS including also stars which have been discovered since the study of Neuhäuser et al. (1995). These newly discovered stars are listed in K"onig et al. (2000). Most of them were not detected during the RASS. We do not include all those TTS, which were originally discovered by ROSAT , in order to avoid a bias towards X-ray bright TTS.
3.3. IC 2602
IC 2602 is a 30 Myr (Mermilliod 1981) old open cluster whose stars are about to reach the main-sequence. With a distance of (see Whiteoak 1961) the cluster is relatively nearby. The X-ray emission from IC 2602 was studied by Randich et al. (1995) who have analyzed pointed PSPC observations and selected probable cluster members on the basis of the photometry of the optical counterparts. For the comparison of the XLDFs we have made use of Table 4 in Randich et al. (1995) (see Sect. 3.5).
Due to their relatively small distance (116 pc; Mermilliod et al. 1997) the old Pleiades cluster has often been used in comparative evolutionary studies. Detailed investigations of the X-ray emission from the Pleiades have been presented e.g. by Stauffer et al. (1994), Micela et al. (1996), and Micela et al. (1999) based on observations performed by the Einstein and ROSAT satellites. A study of the full set of archived ROSAT observations will be presented in a later publication (Stelzer et al., in prep). Here, we anticipate the XLDF composed of all Pleiades stars which have been in the field of view of any pointed PSPC observation. Upper limits for non-detections are included. A detailed description of the data analysis will be given in the subsequent paper.
3.5. X-ray luminosity functions
To compute the XLDFs for the different stellar associations we have used the ASURV statistics package (see Feigelson & Nelson 1985) which ensures a proper treatment of censored data points, i.e upper limits for undetected sources. For unresolved multiples we have assumed that all components emit X-rays at the same level. The observed has, therefore, been divided by the number of components, and all components are considered in the XLDF.
In Fig. 1 we display the RASS XLDF of the group of probable
members of the Tucanae association. The subsample of late-type stars is also shown. For these stars the XLDF is somewhat steeper indicating that the early-type stars are the weaker X-ray emitters. Five of the early-type stars in the Tucanae sample have spectral type A. For A type stars no mechanism producing X-rays is known, consistent with our finding that all A stars in Tucanae are undetected in the RASS. O and B stars can generate X-rays in shocks associated with their strong winds. None of the B stars from our sample is detected in the RASS, but one B star is detected in a PSPC pointing. For late-type stars with convective envelopes of substantial depth (starting from F5; Walter 1983) the X-ray emission is thought to be related to magnetic dynamo activity. All stars in Taurus-Auriga are TTS, i.e. late-type PMS stars. Therefore, to obtain homogeneous samples for the comparison of XLDFs with other star-forming regions only the G, K, and M stars among the Tucanae members should be retained. Due to the non-detection of stars earlier than spectral type F in the Tucanae sample, reducing the sample to G, K, and M stars mainly effects the number of upper limits involved in the XLDF.
We have reduced the other samples in the same way to G, K, and M members. In TW Hydrae most stars have very late spectral types. Only one star, HR 4796A (spectral type A0), had to be excluded. All TTS in Taurus have spectral types G and later. The sample of IC 2602 is composed of all stars from Table 4 in Randich et al. (1995) which are labeled `photometric members' (flags `Y' or `Y?') and have . In this table the authors give for all ROSAT PSPC X-ray sources in the cluster with optical counterparts. To treat the multiples among these stars consistently with the other stellar groups we have made use of the Open Cluster Data Base compiled by C. Prosser and colleagues (available at ftp://cfa-ftp.harvard.edu/pub/stauffer/clusters ). The multiplicities given in the Open Cluster Data Base are used to derive of the individual components as described in Sect. 2.1. The membership list in the Pleiades is based on the entries of the Open Cluster Data Base. For the XLDF computed for comparison with Tucanae we restrict this sample to G, K, and M type stars which have been observed in any pointed PSPC observation (see Stelzer et al., in prep. for more details).
The Kaplan-Meier Estimator for the late-type stars of all stellar groups introduced in the previous subsections is shown in Fig. 2.
All distributions except that of the Pleiades are remarkably similar. Particularly, the XLDF of the 30 Myr old IC 2602 cluster and the PMS regions occupy the same region in the diagram. The distribution for TW Hydrae shows the steepest slope, i.e. smallest spread of luminosities. The narrow luminosity distribution of the TW Hydrae sample might be explained by two effects: (i) Since only for four systems the parallax has been measured, we have adopted a mean distance of 55 pc for the remaining stars. Therefore the real spread in distance is probably underestimated. And (ii) the spectral type distribution is very homogeneous in TW Hydra. Most stars have spectral types late K or M, while for the other samples the spread in spectral types is larger (a substantial number of G and early K stars enter the distributions). We note, that the distribution of the Pleiades and IC 2602 have been derived from pointed data, while the XLDF of Tucanae, TW Hydrae and Taurus-Auriga are obtained from RASS observations. However, in the displayed luminosity range the lower sensitivity limit of the RASS should not play a role, and all distributions should be complete.
The general coincidence of the shape and location of the XLDF of Tucanae, TW Hydrae, Taurus-Auriga, and IC 2602 suggests that the stars in the Tucanae association are young (10 to 30 Myr). Certainly, their age is well below that of the Pleiades () whose XLDF is clearly shifted to lower luminosities. This is also manifest in Table 5 where we give the mean and median of the X-ray luminosity for all examined samples.
Table 5. Mean and median of the X-ray luminosity for the stellar samples compared in Fig. 2 computed with ASURV. Results for all stars (regardless of spectral type), and for the samples restricted to G, K, and M stars are given. The number of stars in each sample and number of upper limits among them are listed in columns `Size' and `(u.l.)'.
The values presented in Table 5 have been derived with ASURV, i.e. upper limits have been taken account of. The weakly X-ray emitting stars of early spectral type in Tucanae reduce significantly. All values for of late-type stars except the Pleiades are compatible with each other within their uncertainties.
The mean ratio of the logarithm of the X-ray to bolometric luminosity for the RASS detected Tucanae members (Table 1a of Zuckerman & Webb 2000) is , which is typical for late-type stars generally characterized by . Four Tucanae stars display an value slightly higher than this saturation limit. But note, that two of these show strong variability (see lightcurves in Fig. 4) which may have led to an overestimation of the quiescent X-ray emission. For the probable members correlates well with the equivalent width of Li I, a commonly accepted age indicator. The correlation is shown in Fig. 3 (filled circles). The stars labeled as `improbable' members by Zuckerman & Webb (2000) do not follow the relation (open circles). In most stars from this group neither the Li I line is detected nor are they known to be X-ray emitters consistent with them not being part of the association.
© European Southern Observatory (ESO) 2000
Online publication: October 2, 2000