A&A 461, 983-990 (2007)
DOI: 10.1051/0004-6361:20065174

Searching for gas-rich disks around T Tauri stars in Lupus[*]

T. A. van Kempen1 - E. F. van Dishoeck1 - C. Brinch1 - M. R. Hogerheijde1

Leiden Observatory, PO Box 9513, 2300 RA Leiden, The Netherlands

Received 9 March 2006 / Accepted 7 June 2006

Abstract
Aims. We characterize the molecular environment of classical T Tauri stars in Lupus observed with the Spitzer Space Telescope in our search for gas-rich disks toward these sources.
Methods. Submillimeter observations of 12CO, 13CO, and C18O in the J=3-2 and 2-1 lines were obtained with the James Clerk Maxwell Telescope toward a sample of 21 T Tauri stars with disks in the Lupus molecular clouds. Pointings at the sources and at selected off-positions are presented in order to disentangle material associated with disks from ambient cloud material.
Results. One source, IM Lup (Sz 82), was found with the double-peaked 12CO and 13CO profiles characteristic of a large rotating gas disk. The inclination of the disk is found to be $\sim$$20^\circ$, with an outer radius of 400-700 AU. For most other sources, including GQ Lup with its substellar companion, the single-dish 12CO lines are dominated by extended cloud emission with a complex velocity structure. No evidence of molecular outflows is found. Due to dense circumstellar material, compact C18O emission was detected toward only two sources. Future searches for gas-rich disks in Lupus should either use interferometers or perform very deep single-dish integrations in dense gas tracers to separate the cloud and disk emission.

Key words: accretion, accretion disks - radiative transfer - methods: observational - stars: pre-main sequence - radio lines: ISM - stars: circumstellar matter

1 Introduction

A general characteristic of pre-main sequence (PMS) stars is the presence of circumstellar disks, which have been observed for a wide range of evolutionary stages associated with low-mass star formation (e.g., Greaves 2005; Lada 1987; André et al. 1993). These disks contain all the necessary ingredients for forming complex planetary systems similar to our own solar system, now also seen around other stars (e.g., Marcy & Butler 1995; Ruden 1999). The gas-rich phase of disks is crucial in the evolution, since gas is needed to build Jovian-type planets. Also, the presence of gas affects the dynamics and evolution of the dust in the system and the migration of any planets embedded in the disk (e.g., Gorti & Hollenbach 2004). With time, the disk will lose its gas through processes such as photoevaporation, stellar winds, formation of gas-rich planets, and accretion onto the star (e.g., Hollenbach et al. 2000; Alexander et al. 2005). However, the timescale for gas dissipation in disks is still poorly constrained observationally, and it is even unclear whether gas and dust are lost simultaneously from disks when they evolve from the massive optically-thick to the tenuous optically-thin "debris'' phase.

To constrain these timescales, surveys of gas in disks in a variety of environments are needed with sensitivities that are sufficient for detecting a Jupiter mass of gas ($\sim$10-3 $M_\odot$). Surveys to date have focussed on CO millimeter observations, mostly of disks in the Taurus molecular cloud (e.g. Thi et al. 2001). Although CO is known to be a poor tracer of the gas mass due to the combined effects of photodissociation in the upper layers and freeze-out in the midplane (e.g., van Zadelhoff et al. 2003), it is much simpler to observe with current instrumentation than other tracers, including H2(Thi et al. 2001) or atomic fine-structure lines (Kamp et al. 2003). Interferometer surveys have detected gas-rich disks around a large fraction (>60%) of classical T Tauri stars in Taurus-Auriga with ages up to a few Myr and gas masses of at least a few times 10-4 $M_\odot$, depending on assumptions about the CO abundance (Dutrey et al. 1996,2003; Koerner & Sargent 1995). In contrast, the detection rate is only 10% for the more evolved weak-line T Tauri stars (Duvert et al. 2000). Similar statistics are found in a single-dish submillimeter survey of CO in a spatially more distributed sample of Herbig Ae and Vega excess stars (Dent et al. 2005) and in CO infrared surveys (Najita et al. 2003; Blake & Boogert 2004).

Surprisingly little is known about the presence of gas-rich disks around T Tauri stars in other nearby star-forming regions such as Corona Australis, Lupus, and Chamaeleon. This is largely due to the absence of interferometers in the Southern sky, and even Ophiuchus has been poorly sampled with current facilities. We present here an initial search for gas-rich disks in the Lupus molecular clouds, using a single-dish telescope.

The Lupus clouds, located at around galactic coordinates 335 $^\circ \le l \le 341^\circ$ and $7^\circ \le b \le 17^\circ$ (Hughes et al. 1994), are among the closest star-forming regions to the Sun at $\sim$150 pc (see reviews by Krautter & Reipurth 1991; Comerón 2006). Mid M-type stars dominate the stellar population of Lupus, which contains no confirmed O or B-type stars. In total 69 pre-main sequence (PMS) objects were found by Schwartz (1977), distributed over four clouds now known as Lupus 1-4 (see also Hughes et al. 1994). In addition, some 130 new weak-line T Tauri stars were detected in ROSAT images (Krautter et al. 1997,1994). The most massive stars in the complex are two A-type stars, HR 5999 and HR 6000, present in the sub-group that is richest in stars and complexity, Lupus 3. A good overview of the clouds is given by the extinction map made by Cambrésy (1999), based on star counts. Recently a higher resolution ($\sim$30'') map was published by Teixeira et al. (2005).

The Lupus clouds have been studied at millimeter wavelengths both in molecular emission lines and in the continuum. 12CO (Tachihara et al. 2001), 13CO (Tachihara et al. 1996) and C18O (Hara et al. 1999) J=1-0 large-scale maps were made at 2 $.\mkern-4mu^\prime$6 resolution with the NANTEN telescope. Higher resolution observations were done with the Swedish-ESO Submillimeter Telescope (SEST) over more limited regions with $\sim$45'' resolution by Gahm et al. (1993) (mostly 12CO 1-0 in Lupus 2), Rizzo et al. (1998) (12CO 1-0 in Lupus 1 and 4 filaments) and Vilas-Boas et al. (2000) (13CO and C18O 1-0 in localized dark cores in Lupus 1-4). Nürnberger et al. (1997) performed a continuum survey of 32 T Tauri stars in Lupus at 1.3 mm with the SEST, probing the cold circumstellar dust.

The Lupus star-forming region appears to differ in several aspects from other low-mass star-forming clouds. In contrast to Ophiuchus and Taurus, which contain a significant fraction of deeply embedded Class 0 and Class I objects, Lupus has at most a few embedded objects (Tachihara et al. 1996; Krautter & Reipurth 1991; Comerón 2006). This could be an indication that Lupus is more evolved than other star-forming regions. The peak in the age distribution is $3\times10^6$ yr for stars in Lupus 3 and 4, calculated from the PMS evolutionary tracks of D'Antona & Mazzitelli (1994) for a distance of 150 pc (Hughes et al. 1994). Under the same assumptions, Lupus 1 and 2 PMS stars are younger with a peak of $1\times10^6$ yr. These age estimates are comparable to those for other star-forming regions such as Chamaeleon (Hartigan 1993), but are about double that of the Taurus cloud (Simon et al. 1993). Thus, a study of disks in Lupus is interesting because this is precisely the age range in which significant evolution of the gas disk is expected.

A second characteristic of the Lupus clouds is the distribution of spectral types. Compared to the Taurus cloud, Lupus is dominated by lower-mass stars. The peak of the distribution in spectral type is M0, with very few stars having spectral types higher than K7. In contrast, the peak of the spectral type distribution in Taurus is at K7.

In this work, single-dish submillimeter observations of CO toward a sample of 21 classical T Tauri stars with disks are presented. They were obtained with a beam of 14'', much smaller than that of previous data. The sources are a subset of the PMS stars found by Schwartz (1977), Krautter & Reipurth (1991), and Hughes et al. (1994), and they have been observed with the Spitzer Space Telescope in the context of the "From Molecular Cores to Planet Forming Disks'' (c2d) Legacy survey (Evans et al. 2003). The aim of the CO observations is to characterize the molecular environment of the T Tauri stars at high angular resolution and to search for gas-rich disks suitable for future follow-up interferometer observations.

2 Observations

The submillimeter CO observations were carried out in two runs in April and July 2005 at the James Clerk Maxwell Telescope (JCMT)[*]. The 12CO J=3-2 and C18O J=3-2 and 2-1 transitions (depending on weather conditions) were observed with the dual-polarization B3 and single-polarization A3 receivers. The beam sizes are 14'' and 20'', respectively. The Digital Autocorrelator Spectrometer was used as the back-end with a bandwidth of 125 MHz, giving a spectral resolution of 98 kHz for the A-receiver corresponding to 0.10 km s-1 at 230 GHz, and 196 kHz for the dual-polarization B-receiver corresponding to 0.22 km s-1 at 345 GHz. The integration times were such that the rms is around 70-100 mK in a 0.2 km s-1 velocity bin for the B-receiver and 50 mK for the A-receiver, depending on source and observing conditions, on a $T_{\rm {mb}}$ scale. Pointing was checked regularly for these southern sources and found to vary within 2''. Beam efficiencies were taken to be 0.65 for RxA3 and 0.63 for RxB3[*].

Frequency switching with a switch of 16.4 MHz was adopted for 12CO because no emission-free position could readily be found for many of the sources in our sample. If a signal was detected on source, additional observations were carried out at 30'' offsets to the east and south of the source to characterize its environment and constrain the on-source contribution. If indications were found of a gas-rich disk, additional follow-up observations were done on-source in 13CO 3-2 using a beam switch of 180'' in AzEl. C18O 3-2 or 2-1 observations were performed for nearly all sources, also using a beam switch of 180''. Table 1 summarizes the observed lines per source. The data were reduced with the SPECX and CLASS reduction packages.

Table 1: Summary of the properties of T Tauri stars and observed lines in Lupus.

Table 2: Observed line intensities toward T Tauri stars in Lupus.

Our observing strategy is similar to that for the single-dish observations of gas-rich disks in Taurus by Thi et al. (2001). Because the surface layers of disks are warm, the 12CO 3-2 line is well-suited to observing these regions. Also, the smaller beam size at higher frequencies is more favorable to observing disks, which have typical sizes of a few hundred AU corresponding to a few arcsec at the distance of Lupus. Typical 12CO 3-2 antenna temperatures of disks in Taurus are 0.5-1 K, so that an rms of <0.1 K needs to be reached. To distinguish gas associated with disks from that of surrounding clouds, observations on-source and at nearby off-source positions need to be taken. Gas-rich disks should show a characteristic double-peaked line profile and a stronger signal on-source than at the off-source positions, while cloud emission should produce a comparable signal at all positions. Finally, disks in Taurus have not been detected in C18O with single-dish telescopes. On the other hand, such beam-switched C18O observations have been found to be a good probe of any dense circumstellar gas or remnant envelope associated with the source. For example, Class I sources show typical C18O intensities of 0.5-2 K (Hogerheijde et al. 1998; Jørgensen et al. 2002).

Our sample of classical T Tauri stars in Lupus was selected from targets belonging to the InfraRed Spectrometer (IRS) sample of the c2d Legacy program on Spitzer (see Table 1). They are a subset of the T Tauri stars studied by Schwartz (1977), Krautter & Reipurth (1991) and Hughes et al. (1994), selected to be brighter than 200 mJy at 12 $\mu$m and excluding sources contained in the Spitzer-IRS guaranteed time programs. Infrared photometry and spectroscopy was performed on all sources giving estimates of their spectral types (see Table 1). In total 21 PMS stars were observed, most of them located in Lupus 3. All sources show a clear 10 $\mu$m emission feature due to warm silicates, confirming the presence of a circumstellar disk (Kessler-Silacci et al. 2006). Some of the selected stars are known binaries. Where available, results of the 1.3 mm continuum emission are reported (Nürnberger et al. 1997). This emission is presumably dominated by the cold dust in the disk.

The distances to individual Lupus clouds have considerable uncertainty. Hughes et al. (1993) found $140 \pm 20$ pc from the spectroscopic parallax of field stars around the clouds. However, observations of reddening and interstellar absorption lines toward individual stars by, e.g., Knude & Høg (1998) and Crawford (2000) find distances ranging from 100 to 200 pc. A good discussion can be found in Comerón (2006). The depth of the Lupus clouds may be non-neglible and can be as large as 50 pc. Also, it has been suggested that not all clouds of Lupus are related; in particular Lupus 2 could be located as far away as 360 pc (Knude & Nielsen 2001). However, this seems unlikely, given the narrow ranges in mean velocity of the clouds (see e.g., Tachihara et al. 1996; Vilas-Boas et al. 2000; and this work Sect. 3). The distance used here for all sources is that most commonly adopted, 150 pc.


  \begin{figure} \par\includegraphics[width=400pt]{12co_br.ps}\end{figure} Figure 1: 12CO 3-2 lines of selected sources in Lupus, with off-positions 30'' south and east of source.
Open with DEXTER

3 Results

The observed 12CO 3-2 line intensities are presented in Table 2, together with the peak intensities. Typical examples of spectra that illustrate the range of observed profiles, both on-source as well as at off-source positions, are shown in Fig. 1. The remainder of the spectra can be found in the on-line material.

Only a single source is found in our sample with a clear signal of gas associated with the source itself, namely IM Lup. A classic double-peaked disk profile is found toward the source position, which is largely absent at the 30'' east and south positions offset from the source (see Fig. 1). The profile, with a maximum peak of 2.5 K, is significantly broader than the one associated with quiescent gas. This source is discussed further in Sect. 4.1.

None of the other sources show indications of gas-rich disks surrounding the central T Tauri star. Three sources, RY Lup, EX Lup and Sz 111, appear to be located outside any cloud material and have no detectable disk emission at our sensitivity limit. Using the formula in Thi et al. (2001) with the observed limits on 12CO emission, upper limits of the gas masses of $1\times10^{-6}$ $M_\odot$ are obtained for these disks with a CO/H2 ratio of 10-4. For a more realistic overall CO/H2 disk abundance of 10-6, the limits would increase to $1\times10^{-4}~M_\odot$ (see also Sect. 4).

Most sources have a strong 12CO line on source, but these signals are also detected at the off-positions with comparable strengths (see Fig. 1). These lines must originate in the large-scale molecular cloud material in the line of sight of the source. The complex profiles seen for many sources indicate a complex structure of different molecular clouds moving at different velocities, either in the fore- or background. For example, the spectra toward Sz 118 (see Fig. 1) show three peaks with different intensities at all positions with widths of $\sim$0.8 km s-1 and velocities that differ by 1-1.5 km s-1. The C18O molecule is not detected at all for this source. Comparison with the large scale maps shows that observations of 12CO 3-2 with peak temperatures of 2-20 Kelvin are consistent with the lower-resolution 12CO surveys (Gahm et al. 1993; Tachihara et al. 1996).

The 12CO line widths of $\sim$0.8 km s-1 are comparable to those found in Vilas-Boas et al. (2000) and Tachihara et al. (2001), taking multiple clouds at different velocities into account. No evidence of line wings at $\geq$2 km s-1 from line center is found at the level $T_{\rm {MB}}<0.15$ K (2$\sigma$), providing limits on any small-scale molecular outflows associated with these T Tauri stars. For RU Lup, this is consistent with Gahm et al. (1993), who reported no outflows from CO 1-0 data in a larger beam.

Although no direct correlation can be inferred due to the confusion of cloud emission around most sources, it is interesting to note that the largest 1.3 mm flux found by Nürnberger et al. (1997) is associated with IM Lup. Thus it also contains the most surrounding dust, in addition to being the only source with CO disk emission.

The results for the C18O and 13CO line intensities are included in Table 2. The 13CO 3-2 molecule was observed for only two objects: IM Lup and GQ Lup. A more detailed description for these two objects can be found in Sect. 4. Figure 2 shows some examples of observed C18O lines.

  \begin{figure} \par\includegraphics[angle=270,width=250pt]{c18o_final.ps}\end{figure} Figure 2: C18O lines toward selected sources in Lupus.
Open with DEXTER

Only two sources in our sample show a C18O line above the noise level, HT Lup and Sz 73, and HT Lup is the only source toward which both C18O 2-1 and 3-2 lines have been detected in our survey. Because the data were taken in 180'' beam-switched mode, the bulk of the emission is likely to be directly associated with the source. This could point towards a small circumstellar envelope. The 2-1/3-2 line ratio of >1 indicates either very cold gas (<15 K) or densities less than 105 cm-3 for $T\approx 20$ K. The inferred C18O column density for these conditions is $\sim$1015 cm-2, corresponding to a CO column density of $5\times 10^{17}$ cm-2 for 16O/18O = 540 (Wilson & Rood 1994) and an H2 column density of $6\times 10^{21}$ cm-2 for a dark cloud C18O abundance of $1.7\times 10^{-7}$ (Frerking et al. 1982). Thus, the extinction directly associated with the source is $A_V\approx 6$ mag using $N_{\rm H}/A_{\rm V}=1.8\times 10^{21}$ cm-2 mag-1 (Rachford et al. 2002).

As mentioned before, our observations are not sensitive to extended C18O emission because of beam-switching by 3'. The lack of any features at the 2$\sigma$ level of <0.1 K in C18O 2-1 and <0.3 K in 3-2 implies that either any large scale cloud emission must be very smooth on a 3' scale or that there is no C18O at this limit. Vilas-Boas et al. (2000) and Hara et al. (1999) find C18O 1-0 antenna temperatures of 0.2 to 1.5 K for the densest cores. For typical dark cloud conditions, the 2-1 line should be comparable in strength, whereas the 3-2 line should be factors of 2-4 weaker, so that these lines would have been detectable at the positions of strongest 1-0 emission. However, most of our sources are located at the edge or outside the C18O contours of Hara et al. (1999), suggesting that C18O is indeed largely absent. The lack of C18O 2-1 emission at the level of 0.08 K ($2\sigma$) indicates a C18O column density of less than $5 \times 10^{13}$ cm-2, or H2 less than $3\times 10^{20}$ cm-2 using the dark cloud abundance. In this diffuse cloud regime, however, the C18O abundance may be significantly lower than the standard dark cloud value (van Dishoeck & Black 1988), increasing the H2 column density.

4 Individual objects

Several objects showed interesting enough spectra to probe deeper. IMLup and GQLup are presented here with a more indepth analysis.

IM Lup is the only T Tauri star in our sample with a clear detection of 12CO emission associated with the source. A double-peaked profile with peak separations of about 2 km s-1 is seen in both 12CO and 13CO, consistent with a rotating gas disk (see Fig. 3). In addition, the lack of C18O emission over 0.09 K in a 0.25 km s-1 bin implies the absence of a significant contribution of a dense circumstellar envelope. The existence of a large disk around IM Lup has recently been shown in scattered light images at visible and near-infrared wavelengths with the Hubble Space Telescope (Padgett et al., priv. comm; Schneider et al., priv. comm.).


  \begin{figure} \par\includegraphics[width=230pt]{imlup_final.ps}\end{figure} Figure 3: 12CO 3-2, 13CO 3-2, and C18O 2-1 spectra observed toward IM Lup (Sz 82) shifted by +4, +2, and 0 K, respectively.
Open with DEXTER

The 12CO peak intensity for IM Lup is comparable to that of the brighter sources with gas-rich disks in Taurus observed by Thi et al. (2001). The much weaker ambient cloud material seen at the south off-position (see Fig. 1) modestly affects the disk profile, however, especially the red peak at $V_{\rm {LSR}}= 5$ km s-1. The blue peak at $V_{\rm {LSR}}= 3$ km s-1 is not contaminated so is taken as the actual disk signal for both peaks. The 13CO peaks of 0.2 K are comparable to those for Taurus disks and do not seem to be affected by ambient cloud material.

The observed 12CO/13CO integrated intensity ratio of $6.0 \pm 1.5$ indicates 12CO and 13CO optical depths of 2.3 and 0.15, respectively, assuming that the excitation temperatures are both the same for both isotopes and constant throughout the disk. The uncertainties are dominated by calibration errors. If the peak temperatures of the blue wing are used (1.5 for 12CO and 0.4 K for 13CO, respectively), optical depths of 3 and 0.3 are found.

Thi et al. (2001) give several methods for deriving disk masses from different observations. Assuming that most of the 13CO 3-2 emission is optically thin and originates in the cold gas of the outer disk with a single excitation temperature, the total gas mass is given by

\begin{displaymath}M_{\rm {gas}} = 3 \times 10^{-6} \bigg( \frac{[^{12}\rm {CO}]... ...~{\rm pc}} \bigg) ^2 \int T_{\rm mb} {\rm d}V \quad M_{\odot}. \end{displaymath} (1)

The H2/CO and 12CO/13CO ratios have been set to 104and 60, following standard conversion factors. This results in a derived gas mass of $2\times 10^{-4}$ $M_\odot$ for IM Lup at d=150 pc, using $T_{\rm {ex}} \simeq 30$ K. For comparison, the disk mass found from the 1.3 mm continuum flux (Nürnberger et al. 1997) is $6\times 10^{-2}$ $M_\odot$, assuming a gas/dust ratio of 100, $T_{\rm {dust}} =30$ K and a dust opacity coefficient $\kappa$ of $0.01~\rm {cm}^{2}~\rm {g}^{-1}$. The factor of 300 between these two numbers is comparable to that found for other disks (e.g., Thi et al. 2001; Dutrey et al. 1996; Mannings & Sargent 2000) and is commonly explained by the fact that CO is not a good tracer of the gas mass due to the combined effects of photodissociation by stellar and interstellar UV radiation and of the freeze-out of CO onto the grains in cold regions of the disk (van Zadelhoff et al. 2001; Aikawa et al. 2002).

The 12CO profile can be used to obtain estimates of two other important disk parameters, the disk size and inclination. Assuming optically thick 12CO 3-2 emission and following the method described in Dutrey et al. (1997):

 \begin{displaymath} \int T_{{\rm mb}} \textrm{d} v \quad=\quad T_{{\rm ex}} (\rh... ...out}-R^2_{\rm in}\right)}{d^2}\cos i \bigg)\quad \Omega^{-1}_a \end{displaymath} (2)

where $R_{\rm {in}}$ and $R_{\rm {out}}$ are the inner and outer radii, $\delta v$ the turbulent velocity, and $\rho$ a geometrical factor. Also, $\Omega^{-1}_a$ is the telescope beam on the sky and i the inclination, with 0 degrees being face-on with respect to the line of sight. For IM Lup, $R_{\rm {in}}= 0 $ AU, a mean disk excitation temperature $T_{\rm {ex}}$ of 25, 40, and 50 K and $ (\rho \delta v)$of 0.3 km s-1 are chosen.

The disk sizes calculated from Eq. (2) are given in $R_{\rm {out}}\sqrt {\cos\ i}$. Scattered light images indicate that the inclination is on the order of 10$^\circ$ to 30$^\circ$, so that $\cos i \approx 1$ is a reasonable approximation. The disk sizes found with $T_{\rm {ex}}=25$ and $T_{\rm {ex}}=40$ K are 730 and 580 AU, respectively. Such numbers are comparable to the observed radius of the dust disk in scattered light, about 4.3'' or 600 AU (Padgett et al., priv. comm.; Schneider et al., priv. comm.).

To better understand the observed line profile of 12CO, a simple disk model has been used, together with a ray-tracing program (Hogerheijde & van der Tak 2000) to compute the line profile. Such models are better-suited to constraining parameters using an $\chi^2$method with the observed spectrum. The populations of the 12CO rotational levels are assumed to be in LTE throughout the entire disk. The surface density of the disk is defined as $\Sigma(r)=\Sigma_0(r/1$ AU)-1.5 with a typical value of $4.57 \times 10^{26}$ cm-2 at 1 AU. The disk is assumed to be isothermal in the vertical direction. The abundance of CO is set to be $5\times10^{-5}$ with respect to H2, except in areas with temperatures below 30 K where CO is frozen out onto the grains. There, abundances are taken to be 5 $\times10^{-10}$. It is assumed that the gas is in Keplerian orbits around a 0.5 $M_\odot$ star. The inner radius is set at 0.1 AU. The radial temperature profile is only constrained by a radial exponent of -0.5 and the luminosity. The turbulent velocity is taken to be 0.1 km s-1, following Thi et al. (2001) and Qi et al. (2004).

The parameters to be fitted are the inclination, the outer radius of the disk, and the temperature at the inner radius. The best-fit is presented in Fig. 4. It has a best fit outer radius between 400 and 700 AU, consistent with the result of 600 AU found above. The line profile is not very sensitive to the outer radius or the CO abundance in the outer areas because temperatures are too low to have a significant impact on the 3-2 emission. The observed line is quite sensitive to the inclination of the disk, which is constrained to $20^\circ \pm 5 ^\circ$. The temperature at the inner radius is taken to be 1200 K, but model results are not sensitive to the precise choice of temperature.


  \begin{figure} \par\includegraphics[angle=270,width=230pt]{bestest_fit.ps}\end{figure} Figure 4: Comparison of the observed 12CO J=3-2 line toward IM Lup with the best-model fit with a disk of $i=20^\circ $, R=600 AU.
Open with DEXTER


  \begin{figure} \par\includegraphics[width=230pt]{gqlup_new_cor.eps}\end{figure} Figure 5: 12CO 3-2, 13CO 3-2, and C18O 2-1 lines toward GQ Lup, shifted by +4, +2, and 0 K, respectively.
Open with DEXTER

The classical T Tauri star GQ Lup has recently attracted attention because adaptive optics imaging has revealed a substellar companion, perhaps a massive Jupiter-like planet, with a separation of $\sim$100 AU (Neuhäuser et al. 2005). As for all objects in our sample, the GQ Lup spectrum shows an infrared excess and a silicate emission feature at 10 $\mu$m, indicating the presence of a dust disk. This object is potentially interesting for our understanding of the influence of a companion on the disk structure and evolution. High signal-to-noise observations were taken toward this object (Fig. 5), but, like most other sources in our sample, only diffuse interstellar cloud material is detected in 12CO, as proven in the observations at off-positions. The 13CO spectrum shows a hint of two peaks centered around $V_{\rm {LSR}}= 5$ km s-1 of equal height at $T_{\rm mb}=0.25$ K, but deeper integrations are needed to confirm this. Also, C18O 2-1 was not detected at all down to 5 mK (2$\sigma$) in a 0.25 km s-1 bin. This constrains the maximum column density of C18O directly associated with the source to $\sim$ $ 2\times10^{12}$ cm-2. An unresolved disk would then be limited to a gas mass of < $2.3\times10^{-5}$ $M_\odot$ for C18O/H $_2=1.67\times10^{-7}$ or < $2.3\times10^{-3}$ $M_\odot$ for a lower ratio of C18O/H $_2=1.67\times10^{-9}$. The mass derived from the 1.3 mm flux of 38 mJy is $2.1\times10^{-3}$, assuming a gas/dust ratio of 100, $T_{\rm {dust}} =30$ K and a dust opacity $\kappa$ of 0.01 cm2 g-1 (Nürnberger et al. 1997).

5 Concluding remarks

The main conclusions of this work are as follows:

- A 12CO J=3-2 survey of 21 classical T Tauri stars with dust disks in Lupus taken in frequency-switched mode has revealed mostly extended diffuse cloud material with a few magnitudes of extinction and a complex velocity structure.

- A C18O J=3-2 and 2-1 survey of the same 21 sources taken in beam-switched mode shows dense compact circumstellar material associated with only 2 sources, HT Lup and Sz 73.

- Only a single source, IM Lup (Sz 82), has been found with clear evidence of a gas-rich disk. Modeling of the line intensity and line profile indicates a large disk with radius $\sim$600 AU and inclination 20$^\circ$. This source will be a prime target for future interferometer observations in the southern sky, especially with the Atacama Large Millimeter Array.
That we do not detect gas disks in Lupus in the current survey does not imply that the gas has been dissipated from these disks; rather, it shows the limitations of single-dish CO line observations for revealing them. Future searches need to either (i) focus on high density single-dish tracers, such as HCO+ J=4-3 or HCN J=4-3 to distinguish the warm, dense gas in the disk from the cold, diffuse cloud material; or (ii) use newly commisioned interferometers, such as the SubMillimeter Array (SMA). From comparison with other disks in Taurus (Greaves 2004; Thi et al. 2004), these lines are expected to have peak intensities of at most 0.1 K in a single-dish 15'' beam, and it will be very difficult to detect. Increased sensitivity to compact emission is only possible with interferometer studies. Interferometric surveys will therefore be essential in determining the fraction of gas-rich disks in Lupus and thus the timescale for gas dissipation in this star-forming region compared with other regions such as Taurus.

Acknowledgements
The authors are grateful to Remo Tilanus for carrying out part of the observations and to Karl Stapelfeldt, Debbie Padgett, Glenn Schneider, and Jean-Charles Augereau for communicating information about the HST images of IM Lup before publication. T.v.K. and astrochemistry at Leiden Observatory are supported by a Spinoza grant and by grant 614.041.004 from The Netherlands Organization for Scientific Research (NWO).

References

 

  
Online Material


  \begin{figure}\par\mbox{\includegraphics[width=140pt]{online/12co/hmlup.ps} \inc... .../sz123.ps} \includegraphics[width=140pt]{online/12co/v908sco.ps} } \end{figure} Figure 6: 12CO 3-2 lines of selected sources in Lupus, with off-positions 30'' south and east of source.
Open with DEXTER


Copyright ESO 2007