YSOs in the Lupus Molecular Clouds

Spectroscopic properties of Young Stellar Objects in the Lupus Molecular Clouds


The results of an optical spectroscopic survey of a sample of young stellar objects (YSOs) and pre-main sequence (PMS) stars in the Lupus Clouds are presented. 92 objects were observed with VLT/FLAMES. All of those objects show IR excess as discovered by the Spitzer Legacy Program “From Molecular Cores to Planet-Forming Disks” (c2d). After reduction, 54 spectra with good signal-to-noise ratio are spectrally classified. Effective temperatures and luminosities are derived for these objects, and used to construct H-R diagrams for the population. The sample consists mostly of M-type stars, with K-type stars. Individual ages and masses are inferred for the objects according to theoretical evolutionary models. The mean population age is found to be between 3.6 and 4.4 Myr, depending on the model, while the mean mass is found to be 0.3 for either model. Together with literature data, the distribution of spectral types is found to be similar to that in Chamaeleon I and IC348. The H line in emission, found in of the sample, is used to distinguish between classical and weak-line T Tauri stars. of the objects show H in emission and are accreting T Tauri stars. Mass accretion rates between and Myr are determined from the full width at of the H peak intensity. These mass accretion rates are, within a large scatter, consistent with the relation found in the literature.

ISM: individual (Lupus) – stars: pre–main sequence – stars: Hertzsprung-Russell diagram –

1 Introduction

Lupus Molecular Clouds is the generic denomination of a loosely connected concentration of dark clouds and low-mass pre-main sequence stars located in the Scorpius-Centaurus OB association at and . Due to its large size, close distance ( = 150–200 pc) and substantial mass of molecular gas, the Lupus Clouds have been subject of many studies at all wavelengths over the years (e.g. Hughes et al. 1994; Merín et al. 2008; see Comerón 2008 for a review; Comerón et al. 2009; Tothill et al. 2009).

Lupus is one of the five clouds selected by the Spitzer Legacy Program “From Molecular Cores to Planet-Forming Disks”, also referred to as c2d (Cores to Disks, Evans et al. 2003). Using 400 hours of observations and all three instruments of Spitzer, the c2d program studies the process of star and planet formation from the earliest stages of molecular cores to the epoch of planet-forming disks. The five observed clouds cover a range of cloud types broad enough to study all modes of low-mass star formation, and large enough to allow statistical conclusions. A rich population of low-mass young stellar objects (YSOs) still surrounded by their circumstellar material has been discovered by Spitzer in these clouds (Harvey et al., 2007a, b; Merín et al., 2008; Evans et al., 2009).

Disks around YSOs, called protoplanetary disks, consist mainly of dust and gas. Dust particles of sub-m size dominate the disk opacity, making it easily observable at infrared and (sub-)millimeter wavelengths by re-emitting some of the received stellar radiation. Dust emission is also temperature dependent with colder dust emitting at longer wavelengths than warm dust. Protoplanetary disks evolve in time, ending up in different scenarios like a planetary system or a debris disk, that may or may not also harbor planets.

The stellar radiation spectrum depends on the effective temperature of the star. Because of the properties of dust emission, the IR excess in a spectral energy distribution (SED) of this star+disk system provides information about the geometry and properties of the dusty disk. To separate star and disk emission, the stellar characteristics need to be known. The inner regions of the circumstellar disk are disrupted by the stellar magnetosphere. The stellar magnetic field creates channels through which material can flow from the disk onto the star. This produces atomic lines from hydrogen, calcium, oxygen, etc. The strongest emission line is H at 6562.8 Å.

The Lupus complex is one of the largest low-mass star forming regions on the sky, with four main star forming sites, referred to as Lupus I-II-III-IV. Furthermore, Lupus III contains one of the richest associations of T Tauri stars (see Comerón 2008 for a review). The work presented here concentrates on data from Lupus I, III and IV because those were the regions observed by the c2d program. Their star formation rates are 4.3, 31 and 4.5 M Myr, respectively (Evans et al., 2009). Distances of pc for Lupus I and IV and pc for Lupus III are assumed in this work (Comerón, 2008).

We present an optical spectroscopic survey designed to characterize the young stellar population of Lupus I, III and IV, as observed by the c2d program. A similar study on a subsample is performed by Alcalá et al. (2011). Their data from an X-shooter survey are still preliminary, but suitable for a parallel study to ours. Section 2 describes the selection criteria for the sample and Section 3 the observations and data reduction. Using libraries of standards from the literature, spectral types and effective temperatures are obtained in Section 4. Combined with additional photometric data, the SEDs of the objects can be built. In Section 5, the stellar and disk luminosities are calculated from the SEDs. In Section 6, the stars are placed in Hertzsprung-Russell diagrams and individual masses and ages are derived based on theoretical tracks. The mass accretion rates, determined through the H line, are presented in Section 7. In Section 8 the young stellar population of Lupus is discussed in context with other regions. Finally, in Section 9, the conclusions from this work are stated.

2 Sample selection

Merín et al. (2008) used the Spitzer c2d point source catalog to identify the YSO population in the Lupus clouds. Objects are classified as YSO if they show an IR excess in the SED. To obtain an optimal separation between young stars, background galaxies and Galactic post-AGB stars, Merín et al. (2008) used the selection criteria developed by the c2d team on its official point source catalog (Evans et al., 2007). The method relies on an empirical probability function that depends on the relative position of any given source in several color-color and color-magnitude diagrams where diffuse boundaries have been determined.

The list of YSOs from Evans et al. (2007) was then adapted by Merín et al. (2008). Visual inspection was performed to subtract suspected galaxies or binaries, leaving the list with 94 YSOs. The final list of Merín et al. (2008) was merged with 65 pre-main sequence (PMS) stars and PMS candidates. Here, the term PMS star is used for other objects added to the list whose youth had already been confirmed using other observational techniques, mainly optical spectroscopy. If an object has not been spectroscopically confirmed as young but it was selected by its optical and near-IR photometry as such, it is labeled as a PMS candidate. This final list of 159 young objects is used for the observations presented here.

3 Observations and data reduction

The data presented here were taken in the second half of the nights of 20 – 25 February 2008 with the Very Large Telescope (VLT) and the instrument FLAMES/GIRAFFE (ID: 080C.0473-A, PI: Oliveira). The instrument was used in the MEDUSA mode, with wavelength coverage of 6437 – 7183 Å, and spectral resolution of 0.79 Å. This wavelength range was chosen for containing temperature sensitive features, useful for spectral classification. Additionally, it covers the H line, an accretion diagnostic. MEDUSA has 135 fibers available, each with an aperture of 1.2″. In total, 250 stars in 19 fields were observed (on average: 14 stars per field). Of those, 158 are field stars, leaving the Lupus science sample with 92 objects. An overview of the observations can be found in Table 1. The different exposure times are adjusted to the mean magnitude of the objects in a given field to avoid saturation of the brightest sources.

The VLT data pipeline, GASGANO, was used for the data reduction. For each observation night, a set of 5 dark frames, 3 flat-fields and 1 arc frame are produced. GASGANO performs bias subtraction, flat-fielding and wavelength calibration. Flux calibration was not required for the science goals since the bands used for classification are close to each other. Further spectral extraction was performed within IDL. In each observed field, unused fibers were placed in random “empty” sky positions. Those positions are not really “empty” sky, since the cloud itself can contribute to the sky with a lot of emission lines. In each field, all the sky spectra are combined within IDL into one master sky spectrum, with a 2 clipping. This combined sky is then subtracted from the science spectra to obtain spectra that are ready for analysis.

The spectra are sorted into four categories: good spectra, non-detection with a mean flux around zero, featureless spectra and not useful spectra (marked as G, U, F and O respectively in Table 1). The two spectra classified as not useful for our science purpose are objects # 14 and 53. Object 14 was discovered to be a galaxy (see section 4.2) while object 53 was observed with a broken fibre, resulting in an unclassifiable spectrum. The non-detections (marked ‘U’, 23 objects) and featureless (marked ‘F’, 13 objects) spectra could also not be spectrally classified. This leads to 54 good classifiable spectra, out of which 26 have a signal-to-noise ratio greater than 20.

4 Spectral Classification

4.1 Method

From the obtained spectra, the YSOs were classified by comparing their spectral features with libraries of standard stars. To match the high resolution of the FLAMES spectra, high resolution standards from Montes (1998) were used. However, standards for all spectral types are not available in this library, so that only a range of spectral types can be derived from it. To further the analysis and narrow down the spectral type, a low resolution (3.58 Å) library was used as well (G. Herczeg, private communication, 2010). Once a spectral type range has been determined from the high resolution standards, the low resolution library is used for a finer determination due to the availability of standards of nearly all spectral types in this library.

The most prominent features in late type stars are the Titanium Oxide (TiO) absorption lines at 7050 – 7150 Å. Mostly K-type and M-type stars have these features because they are cold enough for TiO to exist. The TiO feature at these wavelengths shows three absorption bands, which are deeper for colder objects. The Li i line in absorption at 6707 Å is also a common feature in young stars. This line was detected in some objects (see Table 2), but was not used to spectrally classify the objects.

To compare with the high resolution standards, the science spectra are normalized (divided by the continuum) and overplotted on the normalized standard spectra of different spectral types (see Figure 1). A spectral range is visually determined by the best match for the TiO bands. The method for the low resolution standards is somewhat different because that library is not normalized. First, the resolution of the science spectra is lowered to that of the standard (Å). This is done by convolving the spectrum with a Gaussian profile. The comparison of the spectra with the low resolution standards then happens in two ways. One way consists of scaling the spectra to the standards by anchoring their fluxes at certain wavelengths (6500, 7020 and 7050 Å). An example is shown in Figure 2 (top left, top right and bottom left). In the alternative method a scale factor is determined by first dividing both spectra and then taking the mean of those values, ignoring extreme features (like H emission lines) in the process. This factor is then used to scale the original spectra to the standards (see bottom right panel of Figure 2). Both methods agree very well, producing four different plots for each spectral type within the determined range. The correct spectral type is again visually determined by the best match at the TiO bands. The typical uncertainty in the spectral classification is one sub-class.

Figure 1: Example of the classification method of SST-Lup3-1 with high resolution standards. The black curves represent the high resolution standard models and the orange curves the science spectrum. At the left of each standard its spectral type is indicated. The range of plausible spectral types is determined to be M4 - M8.5
Figure 2: Example of the classification methods of SST-Lup3-1 with the low resolution standard of a M6 main sequence star. The first three panels show the first method where the spectra are anchored at different wavelengths. The last panel shows the second method.

4.2 Special spectra

Some spectra present special features besides the temperature sensitive ones used for spectral classification. These objects were inspected more closely and are differentiated in three special types of spectra:

  • Ten spectra show [He i] emission lines at 6677.6 and 7064.6 Å (objects # 7, 52, 76, 82, 83, 84, 85, 87, 91, 93 - see examples in Figure 3). All of these objects also show strong H in emission. Strongly accreting objects produce emission lines other than H, such as this line. It is concluded that these lines are a sign of accretion, as is H (see Section 7 for more information on this subject).

    Figure 3: Sample of spectra (#82 on left panel and #7 on right panel) with strong H (6562.8 Å), and He i (6677.6 and 7064.6 Å) emission lines.
  • Four spectra (objects #50, 59, 78 and 88) show the [Nii] doublet at 6548.4 and 6583.4 Å and the [Sii] doublet at 6715.8 and 6729.8 Å (see examples in Figure 4). These lines are indicative of ionised gas. Par-Lup3-4 (#50) and Sz102 (#78) are related to Herbig-Haro (HH) objects HH600 and HH228 respectively (Wang & Henning, 2009). They classified these objects to be HHs with the [Sii] doublet line at 6717/6731 Å. For Par-Lup3-4 only these emission lines were detected. This is consistent with Huélamo et al. (2010) who show that this source has a close to edge-on disk. This would make the source appear underluminous.

    The two other objects that show similar spectra, SSTc2dJ160708.6-391407 (#59) and Sz133 (#88), are not yet classified as Herbig-Haro, and are thus designated as Herbig-Haro candidates.

    Figure 4: Sample of spectra (#78 on left panel and #59 on right panel) with [Nii] and [Sii] doublets (at 6548.4 and 6583.4, and 6715.8 and 6729.8 Å respectively). Those sources may be associated with Herbig-Haro objects.
  • One spectrum, IRAS15589-4132 (#14), shows lines at 6757.95, 6774, 6794.8, 6931.5 and 6947.8 Å, and no other features (see Figure 5). These lines are identified as the doublets [N ii] (at 6548.4 and 6583.4 Å) and [S ii] (at 6715.8 and 6729.8 Å) and the strongest H line (at 6562.8 Å) at a redshift . This suggests that the object is a red-shifted galaxy rather than a YSO in the Lupus Cloud. In the NASA/IPAC Extragalactic Database (NED), there is a galaxy classified at the position of this object: 2MASSJ16022165-4140536. For further analysis, this object will no longer be considered part of the sample, leaving 91 sources in total.

    Figure 5: Spectrum of IRAS15589-4132. The emission lines correspond with the spectral lines [N ii] (at 6548.4 and 6583.4 Å) and [S ii] (at 6715.8 and 6729.8 Å). This object is identified as a galaxy at redshift .

4.3 Spectral Types

Following the method described above, the resulting spectral types with their error ranges are listed in Table 2. The objects are mostly M-type stars, but are K-type. The histogram in Figure 6 shows the spectral type distribution, peaking at M6.

40 of these 54 objects were designated by Merín et al. (2008) as YSOs. Spectral classification is found in the literature for 31 of the 54 objects (). The literature classifications match well (within one sub-class) with those derived here. One object, however, differs: object # 26 is classified by López Martí et al. (2005) as M5, but as K0 (range from K0 to K2) here. Their spectral type determination of M5 was based on photometry, which is not as reliable as the spectral classification performed here. For further analysis, the values derived in this work are used.

Figure 6: Distribution of spectral types found in this work. The distribution peaks at M4–M6.

4.4 Effective Temperature

Because spectral types are directly related to effective temperature , the temperatures of the stars classified here can be obtained (see Table 3). For spectral types in the range K – M0, the relationship adopted by Kenyon & Hartmann (1995) was used, while the relationship from Luhman et al. (2003) was adopted for types later than M0. The uncertainties in the temperature are determined directly from the spectral type uncertainties.

5 Spectral Energy Distributions

For a given star with known spectral type (and therefore ), an SED can be constructed using the correct stellar atmosphere model and additional photometric data. The templates for stellar atmospheres used here are the NextGen Model Atmospheres (Hauschildt et al., 1999; Allard et al., 2000). In addition, various photometric datasets are available for the objects in Lupus. Optical and near-IR photometry in the R, I and Z bands (at 0.6517, 0.7838 and 0.9648 m, respectively) was obtained from Comerón et al. (2010, private communication). The data were taken with the Wide Field Imager (WFI) at the La Silla 2.2 m Telescope in Chile. Near-infrared Two Micron All-Sky Survey (2MASS) photometry at J, H and K (at 1.235, 1.662 and 2.159 m, respectively) is publicly available. Finally, Spitzer mid-infrared data at the IRAC and MIPS bands (at 3.6, 4.5, 5.8, 8.0, 24 and 70 m, respectively) are also available online 3 (Evans et al., 2003).

For each object, its photometric data and corresponding NextGen model atmosphere are used (matching its spectral type and the gravitational acceleration for a pre-main-sequence star). The observed photometric data are corrected for extinction at all wavelengths from the object’s visual extinction (), using the extinction law by Weingartner & Draine (2001) with 5.5. This extinction is determined by SED fitting, with considering the derived spectral types. Subsequently, the NextGen atmosphere is scaled to the extinction corrected photometric data at a given band. The scaling band used is either the z or band, depending on the availability of the photometry and the quality of the fit. In Table 2, the extinction values and normalization bands are presented for all fitted objects.

Figure 7 shows the constructed SEDs for 41 out of 54 objects. Here, the solid black line is the NextGen model atmosphere, the open squares are the observed photometry while filled circles show the extinction corrected data. SEDs could not be produced for 13 out of the 54 spectrally classified objects. Either not enough infrared (both 2MASS as Spitzer) photometry is available, or the different photometric data (optical, 2MASS and Spitzer) did not match well with each other. Eight of them are in crowded regions (18, 19, 23, 26, 47, 48, 54, 89), which may have contaminated the photometry. None of the 13 objects are known variable stars.

Figure 7: SEDs of all the objects. The black curve is the model atmosphere of a star at its given temperature. The open points are the observed optical, 2MASS and Spitzer data. The closed points are the extinction corrected data. Notice the wide variety in SEDs (and thus disk shapes).

A wide variety of SEDs can be seen in Figure 7. Consequently, a wide variety in disk types is inferred. The infrared spectral index slope can be used as a characteristic to distinguish between Class I, II and III objects, as defined by Lada. For the 39 objects with available data in K and MIPS1, the slope can be calculated as:

These values are shown in Table 2. One object has consistent with a Class I source () and 15 consistent with Class III sources (), the majority of them are Class II, disk sources. Additionally, object #67 is a confirmed cold disk, as studied by Merín et al. (2010). Cold disks are disks with large inner dust holes (i.e. lack of warm dust). Objects # 27, 79 and 80 have SEDs consistent with inner dust holes. Although confirmation of their nature as cold disks requires SED modelling that is out of the scope of this work, 4 out of 41 objects with SEDs in Figure 7 amount to 10% of cold disks in the sample, which is consistent with the range of 4 – 12% of cold disk frequency derived by Merín et al. (2010).

5.1 Luminosities

Stellar and disk luminosities can be derived for the objects for which SEDs could be plotted. Stellar luminosity is calculated with the formula:

where is the distance to the source and is the stellar flux – an integration of the NextGen model atmosphere scaled and normalized to the dereddened photometric data. Errors are derived from the uncertainty in the distance and flux, which includes an error in of mag. The disk luminosity is the integrated excess emission above the stellar photosphere. Errors are derived from the stellar luminosity errors. Those two values are shown in Table 3.

6 H-R diagram

The age and mass of a star can be derived from theoretical tracks, overlaid on physical H-R diagrams. The position of an object in the H-R diagram is determined by its temperature (derived in Section 4.4) and luminosity (derived in Section 5.1).

6.1 Results

The pre-main sequence evolutionary tracks of Baraffe et al. (1998, 2001) and Siess et al. (2000) are overlaid in the H-R diagrams where the objects in this sample are placed (Figure 8). Thirteen objects are outside the range of the tracks (red and yellow circles in Figure 8).

Figure 8: HR diagram of the objects. Overlaid are the isochrones (black) and mass tracks (blue). The top panel shows the tracks from Baraffe et al. (1998, 2001) and the bottom panel the tracks from Siess et al. (2000). The corresponding ages (in Myr) are marked. The red and yellow circles are objects that do not follow the tracks.

Three objects (marked in red) are below the tracks: ID # 65, 71 and 78. By looking at their SEDs in Figure 7, it can be seen that those objects have substantial IR excess. The stellar fluxes are also not that high. These may be objects still surrounded by envelopes, which could be responsible for their low luminosity. Another explanation is that the objects are seen edge-on, such that the dust in the disk blocks some of the light of the star. In both cases, the calculated luminosity will be too low, resulting in an incorrect placement in the HR diagram. Yet another reason may be that these are background sources and are in fact at a larger distance, which would increase the calculated luminosity.

Ten objects (marked in yellow) are above the tracks: ID 3, 13, 55, 60, 61, 63, 64, 68, 74 and 75. An easy explanation for the misplacement would be that the assumed distance is wrong and that the objects are actually closer to us. However, even with very small distances (e.g. 60 pc), the objects are still above the tracks. Other effects have to play a role. None of them have detected lithium absorption and none of them, except object 13, show any infrared excess. One option is that they are part of a binary system that is unresolved, resulting in a stellar luminosity that is too high. Another option is that they are evolved AGB-stars, which would explain the IR excess seen in object 13. The working assumption is that these objects, however, do not belong to the Lupus Clouds.

Figure 9: Histograms of the ages and masses as derived from the two models.

For those 13 objects, no ages and masses were derived. For the other objects, the effective temperature, stellar and disk luminosity, age and mass can be found Table 3. As can be seen from Figure 9, the two models agree on the range of ages and masses, albeit with a different distribution. The mean age is 3.6 Myr for the Baraffe models and 4.4 Myr for the Siess models. For the masses, both models determine a mean stellar mass of only 0.3 M. This result is in agreement with that derived by Merín et al. (2008). When deriving a luminosity function for the young stellar population in Lupus, the authors found it to peak at 0.2 L, which corresponds to 0.2 M according to the PMS evolutionary tracks of Baraffe et al. (1998).

7 Accretion based on H emission

From the data presented here, 45 of the objects (48 ) show the H line in emission. The strength of the H emission line is often used to distinguish between two classes of pre-MS stars: classical T Tauri stars (CTTS) and weak-line T Tauri stars (WTTS). CTTS are believed to correspond to stars that are actively accreting and WTTS are not, respectively. To this end, different methods have been proposed in the literature. Here, two methods are explored: the equivalent width and the full width of H at of its peak intensity.

7.1 H Equivalent Width

The value of equivalent width (EW) can be used to distinguish between accreting and non-accreting objects, with the dividing threshold depending on the spectral type of the star. White & Basri (2003) proposed that a star is a CTTS if EW[H] 3 Å for K0-K5 stars, EW[H] 10 Å for K7-M2.5 stars, EW[H] 20 Å for M3-M5.5 stars and EW[H] 40 Å for M6-M7.5 stars. Barrado y Navascués & Martín (2003) extended this criterion to the substellar regime. Positive values designate emission. It is important to note that stars with EW[H] values lower than these proposed levels are not necessarily WTTS. Extra confirmation is needed from other diagnostics such as the Lithium abundance.

Following the criterion of White & Basri (2003), 24 of the 45 objects (53.3%) are considered to be actively accreting (values can be found in Table 4). For objects # 22, 49, 50 and 59, no spectral type is available. However, since the EWs of these 4 objects are greater than 40 Å, they are considered accretors anyway.

7.2 Full width of H at of peak intensity

White & Basri (2003) proposed the use of another quantity to distinguish accreting and non-accreting objects: the full width of the line at of the peak intensity, H. Lines with H km s are, according to their measurements, assumed to be accreting. Based on physical reasoning as well as empirical findings, Jayawardhana et al. (2003) adopted 200 km s as a more reasonable accretion cut-off, especially for low-mass stars. This is the cut-off value used in this work.

By performing a Gaussian fit (see Figure 10), H values can be obtained. To account for the resolution, the measured FWHM of each profile must be deconvolved assuming a Gaussian instrumental profile. For this sample, the H values can be found in Table 4. The errors are propagated from the error in the Gaussian fit to the error on the spectral resolution. According to the criterion of Jayawardhana et al. (2003), 25 of the 45 objects (55.5%) are considered accretors. The other 20 objects are not consistent with CTTS, even within their uncertainties. They are classified as non-accretors.

Figure 10: Continuum subtracted, normalized profiles of the H emission lines (black line) overplotted with a Gaussian fit (red solid line). The red dotted line marks the centre of the Gaussian fit and the horizontal red line marks the width at 10% intensity.

Some line profiles show broad wings at the bottom, which cannot be properly fitted by a Gaussian. In order to reproduce those lines, Voigt profiles were fitted to all the lines. In most cases, the fit is either the same or slightly better at the wings as can be see in Figure 11. However, this correction has no significant effect on the full width at . Due to this insignificant difference, only the values from the Gaussian fits are further considered in this work.

Figure 11: Sample of continuum subtracted, normalized profiles of the H emission lines overplotted with a Gaussian fit (red dashed line) and a Voigt fit (thick blue solid line). The red dotted line marks the center of the Gaussian fit.

The EW and H methods are generally consistent in the classification of sources. Only for three objects (#65, 80, 81) do the two methods give discrepant results. This may be the result of the different cut-offs in the EW method, depending on the spectral type.

7.3 Mass accretion rate

The H value can not only be used as an indicator of accretion, but also allows a quantitative estimate of the mass accretion rate. Natta et al. (2004) derived a correlation between H and mass accretion rate (derived from veiling and H profile model fittings):

where is in M yr and H in km s. The relation derived by Natta et al. (2004) was only calibrated for . The cut-off of 200 km s agrees with a cut-off of M yr for . With this relation, mass accretion rates can be calculated for the objects in this sample with H in emission (see Table 4). In Figure 12 the distribution of the mass accretion rates is shown. Typical uncertainties are an order of magnitude. Most of the accretors have mass accretion rates that are consistent with the typical range for classical T Tauri stars (Natta et al., 2004; Sicilia-Aguilar et al., 2010).

Figure 12: Histogram of the mass accretion rates derived from the H value. The shaded area are objects out of the range for which the relationship of Natta et al. (2004) was calibrated. Those objects are generally classified as non-accretors.

A fit of the mass accretion rate as a function of the stellar mass is attempted. This fit reveals that a relation is valid for both models, within the large errors. The same relation, and huge scatter, can be found in the literature (Natta et al., 2005; Alcalá et al., 2011).

8 Discussion

Our work can be added to some of the objects classified as young stars by Merín et al. (2008) that have previously been spectrally classified in the literature in different studies (López Martí et al., 2005; Allers et al., 2006; Allen et al., 2007; Merín et al., 2007; Comerón, 2008). This amounts to 91 of the 159 IR excess sources in the sample of Merín et al. (2008) to have known spectral types. This fairly complete sample can be compared to other nearby star-forming regions with equally complete stellar characterization.

Luhman et al. (2003) studied the young cluster IC348. Spectral types are determined with low-resolution optical spectroscopy for 169 objects. Similarly, Luhman (2007) determined spectral types for the 85 targets in the star forming region Chamaeleon I. Figure 13 shows the distribution of spectral types in Lupus (this work + literature), in IC 348 and Chamaeleon I. The vertical dashed line shows the completeness of the IC 348 and Cha I samples, which is comparable to the limit achieved by Merín et al. (2008). It can be seen that the three regions show very similar distributions of stellar types, from which is inferred that the IMF of Lupus is very similar to those of IC 348 and Cha I.

It is important to note that the sample of Merín et al. (2008) is selected on the basis of IR excess and it is, by definition, biased against young stars without disks (i.e. no IR excess). A similar study characterizing the young stellar population without disks in Lupus is still lacking.

Figure 13: Distribution of spectral types in Lupus (black solid line), IC 348 (green dot-dashed line) and Chamaeleon I (blue dashed line). For Lupus, the distribution includes both our results as well as literature spectral types for sources not in our sample.

When comparing the stellar luminosity with the disk luminosity, a correlation can be found with a correlation coefficient of . The spectral index is also compared with the stellar parameters (mass, age, luminosity). Slight trends can be seen in that the IR excess increases with mass and decreases with the stellar and disk luminosity. However, the scatter is huge in all the fits. More observations and a larger sample are needed to make significant conclusions.

9 Conclusions

We have presented a spectroscopic survey at optical wavelengths designed to determine spectral types and confirm the pre-main sequence nature of a sample of young stellar objects in the Lupus Clouds found in the Spitzer c2d survey.

  • Spectral types were determined for 54 stars belonging to Lupus I, III or IV. The sample consists mostly of M-type stars (90%), but also a few K-type stars. No early type object was found in this sample. The distribution of spectral types peaks at M4–M6. The distribution of spectral types is very similar to that of well studied star-forming regions like IC 348 and Chamaeleon I.

  • The objects were placed in a H-R diagram after effective temperatures and luminosities were derived. Comparison with theoretical isochrones and mass tracks from models of Baraffe et al. (1998) and Siess et al. (2000) yield individual ages and masses for the objects. 10 objects are too luminous to belong to the clouds, while 3 objects are too faint. The very faint objects could still belong to the clouds by having a remnant envelope or an edge-on disk dimming some of the stellar luminosity.

  • The theoretical models by Baraffe et al. (1998) and Siess et al. (2000) imply a population of YSOs concentrated in the age range between 1 and 5 Myr. The mean age is found to be 3.6 and 4.4 Myr with the Baraffe et al. (1998) and Siess et al. (2000) tracks, respectively. Individual masses range from 0.1 to 1.0 M, with mean values of 0.3 M for both models.

  • About half of the sample shows the H line in emission, an important indicator of accretion. This relationship was explored in two different ways: the equivalent width of H, and its full width at 10% of peak intensity. This confirms 25 objects (or 56% of the YSO sample) to be actively accreting objects classified as classical T Tauri stars. The quantitative estimate of the mass accretion rate based on the full width of H at 10% of the peak intensity yields a broad distribution of values ( Myr), typical of T Tauri stars.


Astrochemistry at Leiden is supported by a Spinoza grant from the Netherlands Organization for Scientific Research (NWO) and by the Netherlands Research School for Astronomy (NOVA) grants. This research was based on observations made with ESO Telescopes at the Paranal Observatory under programme ID 080C.0473-A. It also has made use of the SIMBAD database, operated at CDS, Strasbourg, France. The authors would like to acknowledge Greg Herczeg for sharing his low resolution library of standards and for many useful discussions, Bruno Merín and Klaus Pontoppidan for comments on earlier versions of the manuscript, and Loredana Spezzi & Fernando Comerón for providing the optical photometry.

ID Name Cloud RA Dec Observing Date t Mark5
(deg) (deg) (s)
2MASSJ16075475-3915446 Lupus III 2008-02-21 U
2MASSJ16080175-3912316 Lupus III 2008-02-21 U
2MASSJ16080618-3912225 Lupus III 2008-02-21 G
2MASSJ16081497-3857145 Lupus III 2008-02-21 U
2MASSJ16084747-3905087 Lupus III 2008-02-21 U
2MASSJ16085373-3914367 Lupus III 2008-02-21 U
2MASSJ16085529-3848481 Lupus III 2008-02-22 G
2MASSJ16085953-3856275 Lupus III 2008-02-21 U
AKC2006-17 Lupus I 2008-02-23 U
IRACJ16083010-3922592 Lupus III 2008-02-26 U
IRACJ16084679-3902074 Lupus III 2008-02-21 U
IRACJ16093418-3915127 Lupus III 2008-02-26 U
IRAS15567-4141 Lupus IV 2008-02-21 G
IRAS15589-4132 Lupus IV 2008-02-24 O
Lup604s Lupus III 2008-02-21 G
Lup607 Lupus III 2008-02-21 G
Lup608s Lupus III 2008-02-21 G
Lup617 Lupus III 2008-02-21 G
Lup650 Lupus III 2008-02-22 G
Lup654 Lupus III 2008-02-21 F
Lup706 Lupus III 2008-02-26 F
Lup710 Lupus III 2008-02-23 G
Lup714 Lupus III 2008-02-26 G
Lup802s Lupus III 2008-02-26 G
Lup810s Lupus III 2008-02-22 G
Lup818s Lupus III 2008-02-22 G
Lupus3MMS Lupus III 2008-02-24 F
NTO2000-0526.9-5630 Lupus III 2008-02-21 U
NTO2000-0532.1-5616 Lupus III 2008-02-21 U
NTO2000-0536.7-5943 Lupus III 2008-02-24 F
NTO2000-0536.7-5956 Lupus III 2008-02-24 F
NTO2000-0537.4-5653 Lupus III 2008-02-21 U
NTO2000-0540.9-5757 Lupus III 2008-02-21 F
NTO2000-0546.4-5934 Lupus III 2008-02-24 F
NTO2000-0554.9-5651 Lupus III 2008-02-24 F
NTO2000-0558.8-5610 Lupus III 2008-02-24 F
NTO2000-0601.7-5616 Lupus III 2008-02-24 F
NTO2000-0605.1-5606 Lupus III 2008-02-26 U
NTO2000-0605.6-5437 Lupus III 2008-02-21 U
NTO2000-0614.0-5414 Lupus III 2008-02-21 U
NTO2000-0615.6-5616 Lupus III 2008-02-24 U
NTO2000-0615.6-5953 Lupus III 2008-02-24 U
NTO2000-0615.8-5734 Lupus III 2008-02-24 U
NTO2000-0617.7-5641 Lupus III 2008-02-24 U
NTO2000-0619.6-5414 Lupus III 2008-02-21 U
Par-Lup3-1 Lupus III 2008-02-21 G
Par-Lup3-2 Lupus III 2008-02-21 G
Par-Lup3-3 Lupus III 2008-02-21 F
Par-Lup3-4 Lupus III 2008-02-21 F
SST-Lup3-1 Lupus III 2008-02-25 G
SSTc2dJ155925.2-423507 Lupus IV 2008-02-26 G
SSTc2dJ160000.6-422158 Lupus IV 2008-02-26 O
SSTc2dJ160002.4-422216 Lupus IV 2008-02-26 G
SSTc2dJ160111.6-413730 Lupus IV 2008-02-21 G
SSTc2dJ160143.3-413606 Lupus IV 2008-02-21 G
SSTc2dJ160229.9-415111 Lupus IV 2008-02-24 G
SSTc2dJ160703.9-391112 Lupus III 2008-02-21 G
SSTc2dJ160708.6-391407 Lupus III 2008-02-21 F
SSTc2dJ160755.3-390718 Lupus III 2008-02-21 G
SSTc2dJ160803.0-385229 Lupus III 2008-02-24 G
SSTc2dJ160901.4-392512 Lupus III 2008-02-23 G
SSTc2dJ160934.1-391342 Lupus III 2008-02-21 G
SSTc2dJ161000.1-385401 Lupus III 2008-02-22 G
SSTc2dJ161013.1-384617 Lupus III 2008-02-22 G
SSTc2dJ161019.8-383607 Lupus III 2008-02-23 G
SSTc2dJ161029.6-392215 Lupus III 2008-02-23 G
SSTc2dJ161034.5-381450 Lupus III 2008-02-25 G
SSTc2dJ161131.9-381110 Lupus III 2008-02-25 G
SSTc2dJ161144.9-383245 Lupus III 2008-02-23 G
SSTc2dJ161148.7-381758 Lupus III 2008-02-25 G
SSTc2dJ161204.5-380959 Lupus III 2008-02-25 U
SSTc2dJ161211.2-383220 Lupus III 2008-02-22 U
SSTc2dJ161219.6-383742 Lupus III 2008-02-22 G
SSTc2dJ161251.7-384216 Lupus III 2008-02-22 G
Sz100 Lupus III 2008-02-21 G
Sz101 Lupus III 2008-02-21 G
Sz102 Lupus III 2008-02-21 G
Sz103 Lupus III 2008-02-21 G
Sz104 Lupus III 2008-02-21 G
Sz107 Lupus III 2008-02-21 G
Sz108B Lupus III 2008-02-21 G
Sz110 Lupus III 2008-02-21 G
Sz113 Lupus III 2008-02-21 G
Sz114 Lupus III 2008-02-21 G
Sz118 Lupus III 2008-02-22 G
Sz130 Lupus IV 2008-02-21 G
Sz133 Lupus IV 2008-02-24 G
Sz94 Lupus III 2008-02-21 G
Sz96 Lupus III 2008-02-21 G
Sz97 Lupus III 2008-02-21 G
Sz98 Lupus III 2008-02-21 G
Sz99 Lupus III 2008-02-21 G
Table 1: ID, name, cloud, position, observing dates and exposure times of the objects.

ID SpT Range A Band Lithium?
M6.5 M4 - M8.5 J
M3 M2 - M4 Z
M6.5 M4 - M8.5 J
M5.5 M4 - M6 Z yes
M5.5 M4 - M6 Z
M5.5 M4 - M8.5
M6 M4 - M8.5
M3.5 M2 - M4
M4.5 M3 - M6 yes
M6 M4 - M8.5 yes
M5.5 M4 - M6
K0 K0 - K2
M6 M4 - M8.5 Z
M5.5 M4 - M6
M5 M4 - M6
M6 M4 - M8.5 J yes
M4.5 M3 - M6 yes
M3.5 M2 - M4 yes
M8.5 M6 - M9.5 J
M6 M4 - M8.5 J
M7 M4 - M9 J
M5.5 M3 - M6 Z
M9 M8.5 - M9.5 Z
M5.5 M4 - M6 Z
M3.5 M2 - M4 J yes
M7.5 M6 - M9 Z
M6.5 M6 - M8.5 J
M4 M3 - M6 Z
M6 M4 - M8.5 J yes
M5.5 M3 - M6 J yes
M6.5 M6 - M8.5 J
M2.5 M0.5 - M4 J
M6.5 M4 - M8.5 J yes
K2 K0 - K7 J
M8 M6 - M8.5 J
M8.5 M6 - M9.5 J
M4.5 M3 - M6 Z yes
M4.5 M3 - M6 Z yes
K2 K0 - K7 J
M4.5 M3 - M6 Z yes
M5.5 M3 - M6 Z yes
M6.5 M4 - M8.5 Z
M5.5 M3 - M6 Z yes
M3 M0.5 - M4 J yes
M1.5 K7 - M2 Z
M4 M3 - M6 Z yes
K7 K2 - M0 J yes
M2 M0.5 - M4 J yes
K2 K0 - K7 yes
M4.5 M3 - M6
M2 M0 - M3 J yes
M4.5 M3 - M6 Z yes
K5 K2 - M0.5 J yes
M4 M3 - M6 J yes
Table 2: Spectral types, visual extinction (A), normalization bands and the spectral index, calculated between the 2MASS K-band (2 m) and the MIPS1 band (24 m), for the classifiable objects. Objects for which SEDs could not be produced do not have all values. The last column indicates if Lithium was detected.
ID T L L Age Mass Age Mass
(K) (L) (L) (Myr) (M) (Myr) (M)
Table 3: Effective temperature T, stellar and disk luminosity, L and L, age and mass for Baraffe et al. (1998 - 2001) and Siess et al. (2000).

ID EW[H] CTTS? H [] CTTS? log Double peak?
(Å) (km s) (M yr)
yes yes
yes yes yes
yes yes yes
yes yes yes
yes yes yes
yes yes
yes yes yes
yes yes
yes yes yes
yes yes
yes yes
yes yes
yes yes
yes yes yes
yes yes
yes yes yes
yes yes
yes yes yes
yes yes
yes yes yes
yes yes
yes yes
yes yes yes
yes yes
Table 4: H line widths and mass accretion rates. A positive EW value indicates emission.


  1. pagerange: Spectroscopic properties of Young Stellar Objects in the Lupus Molecular Clouds4
  2. pubyear: 2011
  3. htpp://ssc.spitzer.caltech.edu/spitzerdataarchives/
  4. http://ssc.spitzer.caltech.edu/spitzermission/observingprograms/legacy/c2dhistory.html
  5. G means good spectra, U non-detections, F featureless spectra, and O not useful spectra. See § 4.3 for more details.
  6. footnotetext: Sources without ages and masses do not follow the evolutionary tracks.


  1. Alcalá, J. M., et al. 2011, Astronomische Nachrichten, 332, 242
  2. Allard, F., Hauschildt, P. H., Alexander, D. R., Ferguson, J. W., & Tamanai, A. 2000, From Giant Planets to Cool Stars, 212, 127
  3. Allen, P. R., Luhman, K. L., Myers, P. C., Megeath, S. T., Allen, L. E., Hartmann, L., & Fazio, G. G. 2007, ApJ, 655, 1095
  4. Allers, K. N., Kessler-Silacci, J. E., Cieza, L. A., & Jaffe, D. T. 2006, ApJ, 644, 364
  5. Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. H. 1998, A&A, 337, 403
  6. Baraffe, I., Chabrier, G., Allard, F., & Hauschildt, P. 2001, From Darkness to Light: Origin and Evolution of Young Stellar Clusters, 243, 571
  7. Barrado y Navascués, D., & Martín, E. L. 2003, AJ, 126, 2997
  8. Comerón, F. 2008, Handbook of Star Forming Regions, Volume II, 295
  9. Comerón, F., Spezzi, L., & López Martí, B. 2009, A&A, 500, 1045
  10. Evans, N. J., II, et al. 2003, PASP, 115, 965
  11. Evans, N. J., et al. 2007, Final Delivery of Data from the c2d Legacy Project: IRAC and MIPS (Pasadena: SSC)4
  12. Evans, N. J., et al. 2009, ApJS, 181, 321
  13. Harvey, P. M., et al. 2007, ApJ, 663, 1139
  14. Harvey, P., Merín, B., Huard, T. L., Rebull, L. M., Chapman, N., Evans, N. J., II, & Myers, P. C. 2007, ApJ, 663, 1149
  15. Hauschildt, P. H., Allard, F., Ferguson, J., Baron, E., & Alexander, D. R. 1999, ApJ, 525, 871
  16. Huélamo, N., et al. 2010, A&A, 523, A42
  17. Hughes, J., Hartigan, P., Krautter, J., & Kelemen, J. 1994, AJ, 108, 1071
  18. Jayawardhana, R., Mohanty, S., & Basri, G. 2003, ApJ, 592, 282
  19. Kenyon, S. J., & Hartmann, L. 1995, ApJS, 101, 117
  20. López Martí, B., Eislöffel, J., & Mundt, R. 2005, A&A, 440, 139
  21. Luhman, K. L., Stauffer, J. R., Muench, A. A., Rieke, G. H., Lada, E. A., Bouvier, J., & Lada, C. J. 2003, ApJ, 593, 1093
  22. Luhman, K. L. 2007, ApJS, 173, 104
  23. Merín, B., et al. 2007, ApJ, 661, 361
  24. Merín, B., et al. 2008, ApJS, 177, 551
  25. Merín, B., et al. 2010, ApJ, 718, 1200
  26. Montes, D. 1998, Ap&SS, 263, 275
  27. Natta, A., Testi, L., Muzerolle, J., Randich, S., Comerón, F., & Persi, P. 2004, A&A, 424, 603
  28. Natta, A., Testi, L., Randich, S., & Muzerolle, J. 2005, MmSAI, 76, 343
  29. Oliveira, I., et al. 2009, ApJ, 691, 672
  30. Sicilia-Aguilar, A., Henning, T., & Hartmann, L. W. 2010, ApJ, 710, 597
  31. Siess, L., Dufour, E., & Forestini, M. 2000, A&A, 358, 593
  32. Tothill, N. F. H., et al. 2009, ApJS, 185, 98
  33. Wang, H., & Henning, T. 2009, AJ, 138, 1072
  34. Weingartner, J. C., & Draine, B. T. 2001, ApJ, 548, 296
  35. White, R. J., & Basri, G. 2003, ApJ, 582, 1109
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Comments 0
Request comment
The feedback must be of minumum 40 characters
Add comment
Loading ...

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description