Volume 547, November 2012
|Number of page(s)||21|
|Section||Galactic structure, stellar clusters and populations|
|Published online||01 November 2012|
Parameters and measurements obtained in this work for Collinder 69 confirmed members and comparison with previous studies.
In this appendix we provide a more detailed analysis for sources that show peculiarities in the properties studied in Sects. 3 and 4. For consistency we grouped the interesting sources from Sect. 4 following the same subsection scheme.
LOri075: this source has been classified as single-line spectroscopic binary (SB1) by Maxted et al. (2008) (but no binarity sign has been reported in Sacco et al. 2008). According to Maxted et al. (2008), the spectral lines for this star show rotational broadening; the authors compared them to those of a narrow-lined star of similar spectral type and estimated a projected rotational velocity of vsin(i) ~ 65 km s-1. They classified the source as SB1, but also noted that there is an asymmetry in the cross-correlation function (CCF) in the form of a blue-wing, particularly when the measured radial velocity corresponds to a redshift. Therefore they suggested that the fainter component in this binary was detected but unresolved in their spectra.
We detected a double-peaked structure in Hα and Li I in our Magellan/MIKE spectra that made us believe that we had spectroscopically resolved the source. While the origin of double-peak in Li I should be related to binarity, the Hα one could arise from an accreting companion, for example. Additional research on the structure of some photospheric lines marginally confirms this idea. In Fig. A.1 we show the double-peaked structure found in some of the most prominent photospheric lines for this object (given the low temperature of the source, these “most prominent lines” are still very weak). We measured a mean relative velocity of ~45 km s-1 (σ ~ 9 km s-1). We synthesized a 3500 K (log (g) = 4.0) Kurucz spectrum (an effective temperature ~100 K hotter than that estimated for our source, but the coolest temperature for the Kurucz collection) in the region of the Ba λ5535 Å line with the same resolution and a rotational velocity close to that derived by Maxted et al. (2008) (~50 km s-1). We checked that the closest line in the synthetic spectra has a relative velocity of ~110 km s-1, much higher than those measured by us. We show on the right-hand side panel of Fig. A.1 that the relative velocity derived for the photospheric lines does not agree with the one that would be measured from the Hα profile. This fact and the weakness of the lines measured force us to consider the resolution of the binary as tentative. We must note that the environmental Hα component (see Fig. 3) of the region or a possible accreting companion could change the relative velocity of the peaks of this emission line.
Left: double-peaked structure found in photospheric lines for LOri075. Restframe velocity and mean relative velocity of the second peak are indicated with dashed lines. The shaded (blue) rectangle shows the ± 3σ area of the second peak location. Right: Hα profile of the same source. Note how the secondary peak dashed line location (calculated from the photospheric lines) does not agree with the position of the peak (see text for details).
|Open with DEXTER|
Comparisons of the Hα line profiles for different observations of the same two objects. Upper panel: LOri068; one of the objects we suspect esperienced a flare during the Keck/LRIS observations (R ~ 2700). Some asymmetry can be seen in the line even though the resolution of the spectrum is moderate. Lower panel: LOri109, another object suspected to have a flare for which no asymmetry in the line has been found.
|Open with DEXTER|
In our study of the Hα variable sources, we found a subset of five objects for which the criterion from Barrado y Navascués & Martín (2003) applied to spectra taken at different epochs provides contradictory results. While for some measurement of the object taken in one epoch (EW(Hα)1), the source would be classified as accretor, for a different epoch measurement of the same object (EW(Hα)2), the Hα emission could be explained purely in terms of activity.
LOri068 and LOri109 were observed twice during ourcampaigns and LOri068 was also observed by Saccoet al. (2008). Both objects areclassified as diskless sources based on their IRAC slopes, andwhile our Li I measurements agree among themselves (and forLOri068 that the one provided by Saccoet al. 2008), the Hα emission is much more intense in one spectrum for each source than in the others. We believe those spectra were taken while the objects were experiencing a flare. In Fig. A.2 we compare the line profiles of those objects in “steady” and “flared” states, and we show that while for LOri068 we can see some asymmetry in the line profile for the intense emission (which would indicate mass motion), this is not the case for LOri109 (even though for the latter the change in EW is much stronger).
LOri091 cannot be classified with certainty as a variable source. There is one measurement of Hα clearly different from the other two available, but that measurement corresponds to a TWIN spectrum of very poor S/N that translates into a large uncertainty on the continuum, and therefore a very large errorbar in the measurement.
LOri075 is an unresolved (or marginally resolved, see subsection A.1) double system, and therefore variability in the measured Hα is expected.
LOri080 is a puzzling case. We observed the object twice, in 2003 (at Las Campanas) and 2005 (at Calar Alto; see Paper I for a description of the instrumentation used in each case), and both measurements agree within the errors. These measurements place LOri080 at the border of being classified as accreting according to the saturation criterion (see next subsection), although the object shows no infrared excess in the IRAC data. No peculiarity has otherwise been found for the profiles of the lines in either spectra. On the other hand, Sacco et al. (2008) also observed LOri080 with FLAMES and found a significantly lower EW for Hα.
In the process of determining the accretion fraction and their relation with the disk properties, we encountered several particular cases that we describe below.
LOri161: this is the brown dwarf from Fig. 5 that, even though its Hα emission places it well above the saturation criterion, has not been classified as harboring a disk according to its SED. The problem with this very faint source is that it was not detected in IRAC channels three and four (5.8 and 8.0 micron, respectively). Since the sensitivity of these channels is lower than that of one and two (3.6 and 4.5 micron), it could be that this object indeed has a disk that we are not sensitive to and that is undergoing accretion. In that scenario, the estimated accretion rate according to the FW10%(Hα) would be ~1 × 10-10 M⊙/yr, which is much lower than the accretion rate derived for LOri156 (~9.5 × 10-9 M⊙/yr), another brown dwarf with the same spectral type that we discuss below, but that harbors an optically thick disk.
An example of such a disk would be a transitional disk whose excess would be only detectable at longer wavelengths. We checked the new release of the WISE catalog (in the preliminary version the source is not detected) and found a counterpart within 1′′. Unfortunately, although the photometry at the longest wavelengths (~11 and 20 micron) shows a clear excess, these measurements have been classified as “U” (upper limit), and therefore we cannot confirm that this source does harbor a disk.
C69XE-009: an X-ray candidate from Barrado et al. (2011) confirmed spectroscopically as a C69 member in Paper I. This object is at the very limit of the saturation criterion; based on its SED it was classified as a candidate transition disk, but the linear fit to the mid-infrared slope is photospheric. Given that it is clearly an active object (detected in X-rays), and that the disk possibility is based on a very slight excess detected only in one infrared band, we assume that the Hα emission has its origin in chromospheric activity and not in accretion.
C69-IRAC-005: this is the source from Fig. 6 exhibiting the highest accretion rate based on the FW10%(Hα) (log (Ṁ) = −5.56 ± 0.25). The Spitzer/IRAC photometry suggests that it harbors an optically thick disk. This particular source was observed with CAFOS in low-resolution mode, with a wider wavelength coverage than that of the other instruments used (see Paper I). Thus, were able to obtain a different estimate for the accretion rate based on the equivalent width measurement of the components of the CaII triplet (at 8498 Å, 8542 Å and 8662 Å). This emission could be a sign of chromospheric activity too, as in the case of Hα, but the obtained equivalent widths for the triplet are too wide, placing our measurements in the broad-line component of the unresolved line structures that is generally related to accretion (see Comerón et al. 2003; Mohanty et al. 2005). Furthermore, as in Comerón et al. (2003), the CaII triplet line ratios are very close to 1:1:1 (quite different from the 1:9:5 expected ratio for optically thin emission).
Detail of the Hα emission of LOri061 in the CAHA/TWIN spectrum. Note the dependence on the measurement of the full width at 10% of the flux with the pseudo-continuum choice, in particular for the light green and the teal cases (for a complete description of the process to determine the different continuums see Appendix A of Paper I).
|Open with DEXTER|
Upper panel: comparison of accreting (red, very low mass / BD LOri126 and brown dwarfs LOri140 and LOri156) and non-accreting (black, with the same spectral type) members with low-resolution spectra. Note the absence of veiling. Lower panel: detail of the Hα emission of the same comparison.
|Open with DEXTER|
We used the following equations to estimate the accretion rate from the CaII triplet (these equations were derived from the accretion line profile study by Muzerolle et al. 1998 and are discussed in more detail in Comerón et al. 2003): \arraycolsep1.75ptwhere FCaII(λ8542) is the flux in the line, mλ is the magnitude of the star at λ8542, and AV is the visual extinction translated into the wavelength of the line of study using Fitzpatrick (1999) relations. Since the bluest photometric point available for this object is the 2MASS J magnitude, we used the best-fitting model to the SED of the source as a scaling factor to estimate mλ. On the other hand, according to the intrinsic colors by Leggett (1992) and our determination of the spectral type (M3), we found a very low AV value of 0.03 mag (much lower than the average value of 0.36 mag derived for the cluster by Duerr et al. 1982, but neither of them would significantly affect this estimation).
We obtained an accretion rate value of ~3 × 10-7 M⊙/yr, almost an order of magnitude lower than the one obtained based on Hα, which gives us an idea of the caveats of estimating accretion rates from measurements that can be easily contaminated by activity or even by wind contributions.
On the other hand, even though the two estimates differ by such a large factor, this object still seems to be experiencing heavy accretion. We compared its spectrum with that of C69-IRAC-002 (another M3 star, observed with the same setup, harboring a disk but with Hα compatible with pure chromospheric activity and no CaII emission) looking for veiling emission, and no blue excess was found in the source (apart from a marginal excess direcly in the blue edge of the spectra that we think corresponds to an instrumental signature and not to a real excess). This result is not surprising since the wavelength coverage starts at 6200 Å, and veiling in young stars is normally detected at bluer wavelengths. Therefore, we need additional spectroscopic measurements to confirm the presence of veiling in this source.
LOri050, LOri061 and LOri063 are the other sources with more than one estimate of Ṁacc.
LOri050 is a spectroscopic binary according to Sacco et al. (2008) and Maxted et al. (2008); it has been classified as Class II according to its mid-infrared photometry and we obtained two spectra with different instrumentation (see Paper I for details). According to the Hα emission, the object is above the saturation criterion in both cases. The estimated accretion rates for both measurements agree well within the errors (log (Ṁ) = −11.09 ± 0.05, −10.91 ± 0.05). Therefore, we are observing a very interesting system with a total stellar mass of ~0.3 M⊙ and a circumbinary disk that is actively accreting.
For LOri063, on the other hand, the two available accretion rate estimates (from Sacco et al. 2008 and this work) differ by more than an order of magnitude (log (Ṁ) = −10.7 ± 0.3, − 11.87 ± 0.07). LOri063 harbors an optically thick disk according to its IRAC photometry, and the change in the full width at 10% of the flux in Hα is also reflected in the change in EW of the line (>9 Å).
Finally, our measurement of LOri061 does not agree at all with that from Sacco et al. (2008) (two orders of magnitude difference, log (Ṁ) = −10.2 ± 0.3, −7.65 ± 0.05). We believe this difference arises from how sensitive the measurement of the FW10%(Hα) is to the local continuum determination. In Fig. A.3 we illustrate the case graphically. While our automatic procedure (see Appendix A of Paper I for details) identifies a local continuum, the thick light green line, other routines fitting global continuum could base their measurements on the teal line. This difference in the determination of the “real base” of the line yields the large discrepancy in the estimated accretion rate. We must note in any case that among our data on accretors, we do not have other sources where the Hα profile can provoke this confusion in the continuum determination.
LOri126, LOri140 and LOri156 are the three brown dwarfs (LOri126 is directly at the limit between a BD and a very low mass star depending on the method used to estimate its mass, see Paper I) from Fig. 6 exhibiting very high accretion rates (log (Ṁ) = −8.78 ± 0.10, − 8.88 ± 0.14, − 8.02 ± 0.10, respectively). According to their mid-infrared slope, the three targets harbor optically thick disks. And according to their very large Hα equivalent widths, they are well above the saturation criterion.
With these high accretion rates some veiling (due to excess emission from the accretion shock) could be expected in these sources (as is the case for LS-RCrA 1, Barrado y Navascués et al. 2004a). To study this possibility, we selected a non-accreting class III source with very similar spectral type (one half subclass) for each brown dwarf and compared the strength of several TiO molecular bands in both spectra. As can be seen in Fig. A.4, no significant differences are found in the continuum level of any pair of sources. Indeed, in the three cases, rλ, defined as F(λ)excess/F(λ)photosphere is negligible. Whilst for LS-RCrA 1, Barrado y Navascués et al. (2004a) found that rλ varies from ~ 1 to ~ 0.25 for the wavelength range 6200–6750 Å, we found a horizontal slope in this interval. The only cases where a linear horizontal rλ does not work are located at the very edges of the detector, and therefore we can conclude that no veiling is detected in any of the spectra in the studied wavelength range (this does not imply that some veiling cannot be present at bluer wavelengths).
Double peaked structure in the Hα emission of C69-IRAC-007 and C69-IRAC-006.
|Open with DEXTER|
C69-IRAC-006 and C69-IRAC-007: both sources are classified as class II based on their IRAC photometry and show a double-peaked structure of the Hα emission, as can be seen in Fig. A.5. Whilst the sky subtraction for C69-IRAC-007 worked very well, some residual could remain for C69-IRAC-006 (although we see no structure on other, very narrow, “sky lines”).
This double peak is not present either in the other emission line detected in both spectra (He I) or in the absorption lines, which to a certain extent excludes the possibility that these sources are spectroscopic binaries (SB2). We do not have an estimate of the rotational velocities of either given the resolution of the VLT/FLAMES observations (R ~ 8000), but these almost symmetrical double-peak structures in Hα have been reproduced with models for higher mass stars with rapid rotation seen almost pole-on (see Muzerolle et al. 2003 and references therein).
The accretion rate calculated for C69-IRAC-007 is shown in Table 6 since this source fulfills the Barrado y Navascués & Martín (2003) criterion (log (Ṁ) = −7.05 ± 0.02). On the other hand, although the measured Hα equivalent width of C69-IRAC-006 lies well below the saturation criterion of Barrado y Navascués & Martín (2003), the wide FW10% measured (~190 km s-1) places this object very close to the limit of accretors according to Natta et al. 2004). In addition, note the resemblance of the Hα profile of C69-IRAC-006 with that of Cha Hα2, an accreting brown dwarf, modeled in detail (and showing peculiarities attributed to the presence of an outflow) in Natta et al. (2004).
© ESO, 2012
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.