Astron. Astrophys. 352, 574-586 (1999)

3. Parallaxes of TTSs and associated clouds

CTTSs and WTTSs observed by Hipparcos are among the brightest members of their respective classes, although they are some of the faintest stars in the Hipparcos Catalogue. Among the 31 stars with significant parallaxes in Tables 1 and 2, 14 are in the Taurus-Auriga star-forming region, 1 is located in Orion, 6 are associated with the Chameleon and 2 with the Lupus star-forming regions, 3 are within the Scorpius cloud complex, and the other ones are more isolated HAeBe stars and the nearby CTTS TW Hya.

3.1. Data quality

As a first check on data quality, we consider the standard error of the parallax measured by Hipparcos. While the derived parallax accuracy of the sample of young stars measured by Hipparcos depends on the median stellar magnitude, it is typically less than about 2 mas, in agreement with the average accuracy of the Hipparcos Catalogue (cf. Fig. 1). The faintest stars of this sample, BP Tau and DF Tau, are those two CTTSs whose distances appear considerably lower than their associated clouds. The standard errors of their parallaxes is a factor 2 or more larger than standard errors of other stars in the YSO sample ², which should be compared to parallax accuracies derived for the low-luminosity end of the Hipparcos Catalogue. Fig. 1 also displays (lower panel) parallax standard errors of the faintest stars with reliable parallaxes found in the Hipparcos Catalogue, which follows the law expected from photon noise only. It shows that DF Tau's parallax standard error is larger by a factor 1.5 than the standard error of stars with comparable magnitude, while BP Tau does not stand out in this sample. Except for DF Tau, then, the precision of parallaxes for the sample of positively detected YSOs appears to be of a quality consistent with Hipparcos data for stars of comparably low brightness.

Fig. 1. Panel a  Relationship between median Hipparcos magnitude and parallax standard error for the pre-main sequence stars of Table 1. Only positively detected stars with are plotted here. Panel b  Same as Panel a , but for the low-luminosity end of the Hipparcos Catalogue.

3.2. Mean distances of T associations

In each T association, stars were grouped according to their positions and common motion. Mean parallaxes of these groups were then derived from the intermediate data (i.e., the abscissae on the Reference Great Circles of the satellite; cf. ESA 1997), using the method developed by Robichon et al. (1999) for computing the mean astrometric parameters of stellar groups. This method explicitly takes into account the fact that the Hipparcos parameters are correlated within a few square degrees.

3.2.1. Taurus-Auriga

The morphology and distance of the several molecular clouds forming the Taurus-Auriga complex is summarized by Ungerechts & Thaddeus (1987). They adopted a mean distance of 140 pc from previous analyses based on star counts, photometric distances of reflecting nebulae and reddening versus distance diagrams of field stars. Several of the young stars observed by Hipparcos in that region are located in the central part of Taurus, the distance of which was determined by Racine (1968) and Elias (1978) from the photometry of a few bright stars associated with nebulosity. By chance, all the calibrators used by Elias (1978) were observed by Hipparcos, and in Table 3 we give the Hipparcos astrometric parameters of these objects. The weighted average parallax and associated standard deviation of these distance indicators is  mas, corresponding to a distance of pc. This rough average Hipparcos parallax is thus in perfect agreement with previous estimates (cf. Kenyon et al. 1994).

Table 3. Distance indicators for the Taurus cloud

The 17 stars selected in the Taurus-Auriga region were divided into 3 groups of respectively 5, 4 and 5 members. The 3 remaining stars HIP 18117, 20777 and 21852 are too isolated to be included in a group. The astrometric parameters of these stars and their mean values are summarized in Table 4. The term GOF there refers to the goodness-of-fit of the astrometric solution (cf. ESA 1997, p. 112).

Table 4. Astrometric parameters of stars in the Taurus-Aurigae complex. For each group, two mean values, with their standard errors and goodness-of-fit (GOF ), are indicated. The first has been computed using all the stars, while the second considers only single stars (i.e., stars with an empty H59 field).

Group 1 contains stars around the cloud L 1495. Their proper motions are quite similar and reinforce the hypothesis they are part of the same structure. The mean parallax of these 5 stars is  mas and  mas when removing the two VIMs HIP 19762 and 20387. Nevertheless, HIP 20160 (BP Tau), has an individual parallax mas, larger than these mean values. The mean parallax is  mas using only the two single stars HIP 20097 and 20388, which each have an empty H59 field. BP Tau is then at more than 2 from these values. It is unclear whether the star is a member of the group with a diverged parallax or is really closer than the group. In the latter case, it would be an unlikely coincidence that this star would be just in front of a T association with a similar proper motion but a distance less than twice as small. We shall discuss BP Tau in some detail in Sect. 4.

Group 2, in Auriga, contains HIP 22910, 22925, 23143, and 23873. The mean parallax is  mas and  mas when removing HIP 23873 which has a stochastic solution.

Group 3, around T Tau itself, contains HIP 20780, 20782, 20990 to which one could attach HIP 19176 and 20390 which are a few degrees beside them. Unfortunately, only HIP 19176 and 20390 have an empty H59 field. HIP 20780 and 20782 form a two-pointing double system (C in H59) in the Hipparcos Catalogue. HIP 20780 has a reliable solution while HIP 20782 is 3 magnitudes fainter and has a very uncertain solution. HIP 20990 is a faint VIM with an unreliable solution. The mean parallax of these 5 stars is mas. It is  mas using the three reliable stars HIP 19176, 20390 and 20780 and  mas using only the two single stars HIP 19176 and 20390.

Putting the 8 single stars (H59 empty) together, we obtain a mean parallax of the Taurus-Auriga complex of  mas corresponding to a distance of  pc. Using the most reliable parallax value for each group (those obtained with the single stars only), the three groups are respectively at , and  pc. These values, although statistically in agreement, could reflect real distance differences of about the angular size of the complex.

3.2.2. Orion

Orion is a very large complex of molecular clouds with several distinct regions (cf. Maddalena et al. 1986). The complex has practically no tangential reflex solar motion so that YSOs are impossible to separate from field stars using astrometric parameters. The only star with detected (marginally significant) parallax in the Table 1 sample is GW Ori. It is associated with molecular clouds surrounding the HII region excited by the O8 star Ori, HIP 26207, at a distance of  pc; and the derived distance of GW Ori is consistent with this value. The star CO Ori is also in the same vicinity, but it is fainter with parallax standard error comparable to its expected parallax.

In addition to the stars listed in Table 1, 15 presumably young Orion stars are found in the Hipparcos Catalogue: the HAeBe stars HIP 24552, 25299, 25546, 26594, 26752 and 27059 (from van den Ancker et al. 1997), and 9 stars detected by ROSAT in the Orion Nebula cluster (Gagne et al. 1995) - HIP 26220, 26221, 26224, 26233, 26234, 26235, 26237, 26257 and 26258. Only the last of these stars has a (marginally) significant parallax. Two other Orion stars detected by Rosat, HIP 26081 and 26926 are foreground stars according to their Hipparcos parallaxes: respectively  mas and  mas. The Orion nebula cluster is probably not a bound cluster but part of a 80 pc long structure connected with the Orion OB1 association (Tian et al. 1996). Among the Hipparcos-detected stars in this cluster, only HIP 26258 has an empty H59 field. HIP 26220, 26221 and 26224 are three components of a quadruple system, HIP 26233, 26234, 26235 and 26257 are also flagged C in H59, while HIP 26237 has a stochastic solution. No mean parallax could therefore be determined for these objects. HAeBe stars outside the Orion nebula cluster are more isolated and cannot be grouped to compute mean parallaxes.

We can nevertheless derive a new, post-Hipparcos estimate of the distance to the Orion A and B clouds by considering the distance indicators of Racine (1968), which are members of the Orion R1 and R2 associations. Among the 8 stars studied by Racine, 5 were observed by Hipparcos. Their parallaxes are given in Table 5. All these stars are single with empty H59 and H61 fields, except for HD 38087, a double with published component solution. The mean weighted parallax of the Orion distance indicators, to which we add the star HIP 26258 discussed above, is  mas. This corresponds to a distance of  pc, to be compared to the value of  pc derived by Racine.

Table 5. Distance indicators for the Orion R1 and R2 associations (Racine 1968)

3.2.3. Chamaeleon

The Chamaeleon, Lupus, and Scorpius clouds have spatial velocities close to that of the Sco OB2 (Scorpius-Centaurus-Lupus-Crux) association, around  km s^-1. The histories of these SFRs are certainly connected even if their present ages, motions and distances are not exactly the same.

The distance of Chamaeleon clouds has been a topic of controversy for a long time; earlier estimates ranged from 115 to 215 pc for Cha 1, and from 115 to 400 pc for Cha 2. From a recent re-investigation of the reddening with distance over a large area around the clouds, Whittet et al. (1997) conclude that the most probable distance of Cha 1 is  pc, and that of Cha 2 is  pc. A similar analysis using Hipparcos distances gives essentially the same result (Knude & Hog 1998).

We can use 10 stars to compute mean parallaxes of groups in the Chamaeleon region. In Cha 1 there are four CTTSs: HIP 53691, 54365 and the pair HIP 54744 and 54738. The two Herbig Ae/Be stars HIP 54413 and HIP 54557 are usually associated with Cha 1 because of their reflecting nebulosities. Inclusion of HIP 54257 is more speculative, since it is a B star not known as a Herbig star but detected by ROSAT. On the basis of its Hipparcos proper motion and parallax, Terranegra et al. (1999) believe that it is a probable member although Gry et al. (1998) place it just behind the Cha 1 complex. The mean parallax of these 7 stars is  mas, whereas it is  mas when considering only the four stars with an empty H59 field, corresponding to our best estimate of  pc for the Cha 1 distance. Note that the group mean parallax is only mas when also removing the possibly dubious member HIP 54257, which rules out that this star is located behind the cloud, since its individual parallax is  mas.

The other subgroups of the Chamaeleon region have few Hipparcos stars. There is the WTTS HIP 58285 in Cha 3, the B star HIP 57192 just in the head of DC 300-17, and the HAeBe star HIP 58520 at less than 1 degree from the head of DC 300-17 but connected to Cha 3 in the literature. In addition, the B star HIP 57192 is placed just behind the DC 300-17 cloud by Gry et al. (1998). Its parallax is  mas, in perfect agreement with the mean parallax of Cha 1. HIP 58520 ( mas) is definitely closer than the Cha 1 cloud. HIP 58285 is faint with a large parallax error () but its astrometric parameters are closer to those of HIP 58520 than to those of HIP 57192. HIP 58520 and 58285 are members of a presumed moving group of young stars described in Frink et al. (1998) and Terranegra et al. (1999). Two other Hipparcos WTTSs belong to this proposed group, the reality of which is difficult to demonstrate. Its spatial velocity is small and is noticed mainly because of its reflex motion with regard to the Sun. Note that one can find many stars with the same motion in the Hipparcos Catalogue, in a distance range of 50-200 pc, which were not detected by Rosat. While some of them may form a moving group with the WTTSs mentioned above, it is also conceivable that this apparent group is an artifact caused by the limiting magnitudes of Hipparcos and ROSAT.

To summarize this section, the only firm evidence concerning the Chamaeleon complex is that Cha 1 is at a distance of about 170 pc, in agreement with recent estimates by Whittet et al. (1997) and Knude & Hog (1998).

3.2.4. Lupus

Re-assessing association membership from Hipparcos results, de Zeeuw et al. (1999) find a mean distance of 140 pc for the Upper Centaurus Lupus association. From the angular extent of the association (about ), and assuming that it is nearly spherical, the distances of individual members are expected to range from about 110 to 190 pc. Two positively detected stars of our sample (RY Lup and V856 Sco) are apparently associated with clouds of the Lupus star-forming region, and their distances are compatible with the above range.

Six stars connected with the Lupus complex can be used for computing mean parallaxes: HIP 77157, 78094, 78317, 79080, 79081 and 78053. HIP 77157, 78094 and 78317 are TTSs in Lupus 1, 2 and 4 respectively. HIP 79080 and 79081 are the brightest stars of a quadruple system (at least) in Lup 3. HIP 79080 is a Herbig Ae star and 79081 a peculiar B star. HIP 78053 is a WTTS in Lupus 3 with a poor Hipparcos solution and a flag C in field H59. We don't include six other WTTSs discovered by ROSAT (Krautter et al. 1997and Wichmann et al. 1997) and discussed by Neuhäuser & Brandner (1998) because they have proper motions not exactly in agreement with CTTSs, so that it is quite impossible to decide whether they are members of the T association, the OB association or are young field stars instead.

The mean parallax computed for the 5 stars with an empty H59 field is  mas, a small value compared to the previous estimates quoted above. However, it results mainly from the brightest stars HIP 79080 and 79081, which have small errors on their parallaxes. If we remove these stars and consider only HIP 77157, 78094 and 78317, we obtain a larger value of  mas - in agreement with previous determinations of the Lupus SFR distance, but with a large uncertainty - and a value of mas for HIP 79080 and 79081. The explanation could be either that HIP 79080 and 79081 are not members of Lupus 3 or that this cloud is farther away than the other associations.

3.2.5. Scorpius

This is another association studied by de Zeeuw et al. (1999), who find a mean distance of 145 pc, in agreement with previous work by Racine (1968). The parallax distribution width is approximately 1 mas, corresponding to distances from 127 to 170 pc. While the parallax of AK Sco is consistent with this result, the fainter CTTS V1121 Oph may be located somewhat closer to us. Note that the quadruple system Oph ³, associated with the dense parts of the cloud where vigorous star formation is currently taking place, is at a distance of  pc, i.e., on the front side of the Upper Scorpius association.

3.2.6. The TW Hya association

TW Hya, the closest CTTS, has long been considered to be isolated but is now known to be part of the closest T association (cf. de la Reza et al. 1989, Gregorio-Hetem et al. 1992, Kastner et al. 1997, Jensen et al. 1998, Sterzik et al. 1999and Webb et al. 1999). The most recent list of members reports 13 pre-main sequence systems in the vicinity of TW Hya for a total of at least 20 objects ranging from an A star to a possible brown dwarf (Sterzik et al. 1999).

Four of these objects are in the Hipparcos Catalogue: HIP 53911 (TW Hya itself) and the three objects given in Table 6. Their parallaxes lead to a depth of 20 pc, comparable to the angular size of the association. We didn't compute a mean parallax because the method is not suited to such a nearby group where the depth is not small in comparison to the formal parallax errors.

Table 6. Hipparcos-detected members of the TW Hya association)

3.2.7. Conclusion

In most cases, we confirmed that the Hipparcos distances to young solar-type stars are comparable to the distances of molecular clouds and/or OB stars with which they appear associated, as anticipated from a large body of earlier work. There are, however, two discrepant individual cases: BP Tau and DF Tau. These are among the most investigated CTTSs, so one may argue that a good part of what we know about the T Tauri phenomenon is based on these stars. On the other hand, there are a few CTTSs that are found outside of molecular clouds, such as RW Aur or, as seen above, TW Hya. Perhaps such cases are not so rare? Before we start to speculate, it appears quite important to convince ourselves that the Hipparcos results for BP Tau and DF Tau are valid. A possible bias can be due to undetected binarity, and we shall examine in the next section the evidence for binarity in Hipparcos data for YSOs.

© European Southern Observatory (ESO) 1999

Online publication: December 2, 1999
helpdesk.link@springer.de