ABSTRACT
We present high-resolution, near-infrared spectro-astrometric (SA) data for the T Tauri star DR Tau using NIRSPEC at the Keck II telescope. Spectro-astrometry obtains sub-seeing spatial information from emission lines originating in a non-point source object, such as a circumstellar disk. We report the first detection of water SA signatures in a protoplanetary disk. Three water features near 3 μm were averaged together to produce the total signal analyzed. Using a disk model, we constrained the position angle of the disk (∼140°), the inclination of the disk (∼13°), and the emitting region of the emission lines (∼0.056–0.38 AU).
Export citation and abstract BibTeX RIS
1. INTRODUCTION
T Tauri stars, with a mass similar to the Sun, are the closest analogs we have to the early solar system. The disks that form around these stars may hold clues to planet formation processes and timescales. One important question is how and when terrestrial planets such as Earth acquired water and organic volatiles. A commonly invoked scenario suggests that volatile-rich minor bodies (i.e., asteroids and comets) originating ≳3–5 AU from the young Sun may have been transported to the inner protoplanetary disk (Morbidelli et al. 2000; Raymond et al. 2004) where some impacted terrestrial planets. Studying the molecular gas in the inner (<5 AU) regions of disks is key to understanding the formation of terrestrial planets and the transport of volatile material in disks.
While modeling spectral emission features can give information about the location of molecules, such models generally produce non-unique solutions. Spectro-astrometry is a technique that utilizes high-resolution spectrometers with excellent imaging quality to obtain sub-seeing and sub-diffraction limited spatial information from emission lines originating in a non-uniform object, such as a circumstellar disk (Beckers 1982; Christy et al. 1983; Aime et al. 1988; Bailey 1998; Acke et al. 2005). A spectro-astrometric (SA) signal results when an emission line spatial profile differs from that of the continuum produced by a point source, measuring the relative offset of the line emission with respect to the nearby (in wavelength) continuum. For a symmetric disk, a slit aligned perpendicular to the disk major axis will not result in a signal, while the signal intensity will be maximized when the slit is aligned on the sky parallel to the disk major axis. Similarly for disk inclination, a face on disk gives no signal and an edge on disk would maximize the signal. Using this technique, it is possible to constrain the spatial distribution of an emitting molecule on sub-AU scales at the distance of DR Tau (140 pc), as well as the position angle (P.A.) of the disk on the sky and its inclination (Pontoppidan et al. 2008, 2011; Whelan & Garcia 2008; Brittain et al. 2009; van der Plas et al. 2009). Pontoppidan et al. (2011) reported recent SA observations of the fundamental band of CO in DR Tau in the M band.
We present high-resolution, near-infrared, SA observations of DR Tau, centered near 3 μm. DR Tau is relatively bright in the near-infrared and exhibits a rich emission line spectrum (Salyk et al. 2008; Mandell et al. 2012). We detected many water and OH emission features originating in the disk. Here we report our SA analysis for water lines.
2. OBSERVATIONS AND SA ANALYSIS
2.1. Observations
High-resolution (λ/Δλ ∼ 25,000), near-infrared (NIR) spectroscopic data from the T Tauri star DR Tau were obtained on 2011 February 16–18 using NIRSPEC at the Keck II telescope (McLean et al. 1998). The observations range from 2840 cm−1 (3.52 μm) to 3455 cm−1 (2.89 μm). The spectral range was chosen to cover many transitions of both OH and H2O. Observations were performed with a 3 pixel (0
43) wide slit, and we used a standard ABBA nod pattern, with a 12'' beam separation along the 24'' long slit. Combining the spectra of the two beams as (A−B−B+A) in a Taylor Series expansion about the mean airmass cancels, to second order, emissions from "sky" lines and thermal background. Observations were optimized for spectro-astrometry, viewing the source at three different P.A.s and their 180° offsets (Brannigan et al. 2006). Table 1 contains the observing information.
Table 1. Observing Log
| Date | Setting | P.A. | Int. Time |
|---|---|---|---|
| (°) | (minutes) | ||
| 16 Feb 2011 | KL2 | −70 | 28 |
| −40 | 16 | ||
| −10 | 16 | ||
| +50 | 16 | ||
| +110 | 16 | ||
| +140 | 16 | ||
| +170 | 20 | ||
| 17 Feb 2011 | KL1 | −70 | 16 |
| −40 | 20 | ||
| −10 | 20 | ||
| +110 | 16 | ||
| +140 | 16 | ||
| 18 Feb 2011 | KL1 | +50 | 16 |
| +170 | 16 |
Download table as: ASCIITypeset image
Data were dark subtracted, flat fielded, and cleaned of high dark current pixels and cosmic ray hits. The orders were then straightened in the spectral and spatial directions. Detailed descriptions of the reduction procedures can be found in Bonev (2005) and DiSanti et al. (2001), and references therein.
2.2. Spectro-astrometric Analysis
The projected velocity of gas in a rotating disk varies as a function of orbital phase and distance from the star. Regions of common projected velocity space map in loop-like structures. Thus the point spread function (PSF) of each velocity channel of a resolved emission line formed in the disk is offset from the position of the star. While the typical image quality of standard NIR observations is ∼0.5 pixel, SA observations can determine the centroid of the PSF to ∼0.005–0.010 pixel. At the distance of DR Tau (∼140 pc), this provides spatial resolution <1 AU.
Once the beam centers were determined, the centroid for each pixel in frequency space was calculated; this "center-of-light" calculation is analogous to a center-of-mass calculation. We adopt the method of Pontoppidan et al. (2008), calculating the centroid as follows:
where C is a correction factor based on the amount of flux not included in the window over which the signal is calculated and is a number of the order of unity. Xv is the spatial offset (the SA signal; see Figure 1), xi is the position of a given pixel, xo is the centroid of the star, and Fi is the flux for pixel i at a specific wavenumber. This signal is diluted by the strength of the feature relative to the continuum. The undiluted signal is calculated as follows:
The spatial offsets for each beam were averaged together to produce the signal for each straightened frame. The signals for each frame were then averaged to create the signal for a given P.A. Each signal was combined with its 180° offset. Pairwise subtracting of the parallel and antiparallel SA signals removes artifacts from reduction and instrumental effects without removing any real signal, i.e., (X0–X180)/2.
Figure 1. The top panel shows the average emission profile in velocity space of three strong H2O 001–000 vibrational transition lines (centered at 3321.56 cm−1, 3415.51 cm−1, 3426.19 cm−1) that generate spectro-astrometric signals. The bottom three panels are the averaged SA signals (Xv) for those features (black line) for each slit P.A. and 180° offset with the best-fit model over plotted (red line). The horizontal axis is in velocity space (km s−1). 1σ error bars are shown.
Download figure:
Standard image High-resolution image3. MODELING THE SPECTRO-ASTROMETRIC SIGNAL
Emission from a disk in Keplerian motion is expected to produce an antisymmetric SA signal, with the blue- and redshifted sides of the line offset in opposite directions. The H2O position–velocity diagrams of the averaged feature and signals for DR Tau are shown in Figure 1. The observed lines were resolved beyond the instrument profile and were identified via the HITRAN database (Rothman et al. 2009).
In order to interpret the observed SA signal and determine the spatial extent of water emission, we modeled a disk that assumes a flat, circular slab rotating with simple Keplerian motion. An adaptive mesh was used to divide the modeled disk into pieces such that any radial step or angular step moves no more than a set model resolution in velocity space. Each of these calculated regions was treated as its own emitting blackbody. Mandell et al. (2012) found that an exponential drop off in flux (∝ R−3) provided a better fit to the OH emission profile of T Tauri star RU Lup. Such an exponential decrease in flux is also consistent with the outer radius limit for water and OH (∼0.5 AU) determined by Pontoppidan et al. (2011) for AS 205N. For this reason, we adopted a flux of R−β with a β value of −3 for our modeling. A constant radial surface density was assumed.
Figure 2 shows a contour plot of the line of sight velocity of the best-fit model disk projected on the sky. From the model, a synthetic spectral feature and its corresponding SA signal was generated, shown as the red lines in Figure 1.
Figure 2. A projection of the modeled disk (as discussed in Section 4) on the sky, the area of which is the best fit emitting range. The contours are plotted for the line of sight velocity calculated by the disk model.
Download figure:
Standard image High-resolution image
The P.A., inclination, and the inner and outer emitting regions were varied to find the signal that best matched the data using a least chi-squared fit. For this work, we kept the flux dependence on radius and mass fixed to the values in Table 2. The P.A. and inclination have a direct impact on the SA signal itself, as discussed in Section 1, while the radius affects the flux term used to calculate the SA signal. The inner and outer radii that generate the best fit correspond to the spatial extent of emission for that feature. Table 2 shows a summary of the fixed and best-fit model parameters. We found that for the disk around DR Tau, the SA signal indicates that strong gas phase water emission originates in a region extending from 0.056 to 0.38 AU for a nearly face on disk (i ∼ 13°).
Table 2. Summary of Model Parameters
| Betaa | P.A.b | Inc.c | Massd | Rin | Rout |
|---|---|---|---|---|---|
| (°) | (°) | (M☉) | (AU) | (AU) | |
| 3 | 140 | 13 | 0.4 | 0.056 | 0.38 |
Notes. aMandell et al. (2012). bP.A. is degrees east of north. cInc. is such that 0° is face on, 90° is edge on. dIsella et al. (2009).
Download table as: ASCIITypeset image
4. DISCUSSION
Spectro-astrometry is a powerful technique for constraining both the disk geometry and the location of molecular emission in a proto-planetary disk. The reported geometries for the DR Tau disk are highly uncertain. P.A.s of 160° (Akeson et al. 2005), 170° (Andrews & Williams 2007), 98° (Isella et al. 2009), and 0° (Pontoppidan et al. 2011) have been cited, with our best-fit P.A. of 140° falling within those values. Literature values for inclination also vary substantially, ranging from nearly face on (9°; Pontoppidan et al. 2011) to nearly edge on (69°; Muzerolle et al. 2003) as well as intermediate values (36°; Isella et al. 2009). We found that a nearly face on orientation with an inclination of ∼13° was required to fit the SA signal, in agreement with the inclination inferred by Pontoppidan et al. (2011) from SA analysis of CO. The nearly face-on orientation of DR Tau is also consistent with the uncertainty in P.A. and could also explain the single peaked emission features that have challenged interpretation of spectra (see, for example, Mandell et al. 2012).
Our SA analysis suggests a water emitting region close to the star, with best fit values for the inner and outer emitting regions of 0.056 and 0.38 AU, respectively. While Pontoppidan et al. (2011) detected a prominent CO SA signal for the T Tauri star AS 205N and a possible weak signal for CO in DR Tau, they did not detect SA signals for either water or OH. An upper limit for the water emitting region in the AS 205N disk of ≲0.5 AU was determined, corresponding to an emitting area of ≲0.8 AU2. This value is consistent with the water emitting area (0.4 AU2) inferred by Salyk et al. (2008) for both AS 205N and DR Tau from modeling the spectroscopic emission features. This also agrees with our calculated emitting area of 0.44 AU2. In addition, it is interesting to note that the inner water emission radius we calculated is similar to the dust sublimation radius in DR Tau (Akeson et al. 2005). Salyk et al. (2011) also found that the inner radii for CO emission in T Tauri stars, including DR Tau, were similar to the dust sublimation radii. This would be expected if the inner emission radius is due to dust shielding volatiles from photodissociation.
Are these values consistent with what would be expected from astrochemical modeling of protoplanetary disks? Water emission is expected to originate from a warm layer in the disk atmosphere (see models by Willacy & Woods 2009; Walsh et al. 2010; Najita et al. 2011; Heinzeller et al. 2011). While these models differ in their details, they are still useful for guiding interpretation of our measurements, especially when coupled with other spectroscopic measurements for DR Tau. Salyk et al. (2008) and Mandell et al. (2012) both modeled emission lines of water in DR Tau and found a gas temperature of ∼900–1000 K. This temperature is typical of terrestrial planet forming regions and is consistent with the temperature of the disk where Walsh et al. (2010) and Willacy & Woods (2009) predict a peak water fractional abundance in the disk atmosphere. In the case of Walsh et al. (2010), however, the water fractional abundance is still high beyond 10 AU. We note that many models do not include mechanisms that may transport material from above the disk midplane beyond ∼2–3 AU to the midplane where it can condense onto icy grain mantles (beyond the "snow line"). This effect is illustrated in Heinzeller et al. (2011) where radial viscous accretion, vertical turbulent mixing, and vertical disk winds were included, and a dramatic decrease in water fractional abundance in the disk atmosphere was predicted. More work must be done to model the complex chemistry and transport mechanisms that are expected to occur in protoplanetary disks, but our results suggest that most of the water emission is originating in the hot inner disk region.
While the model fits to the water SA signals in DR Tau (Figure 1) reproduce the height of the feature, there is a slight (∼3 km s−1) blueshift of the signal relative to the model. One possible explanation could be that we are sampling a slow disk wind. We note that the heliocentric velocity for the water emission features (26.5 km s−1) is similar to that for stellar absorption lines reported by Alencar & Basri (2000) (27.6 ± 2 km s−1), but given the uncertainty, a slow disk wind, such as that suggested by Pontoppidan et al. (2011) to explain the CO SA signals, cannot be excluded. We will investigate this possibility in future modeling. Another possibility is that a strongly asymmetric dust distribution could lead to a slight offset of the near-infrared continuum centroid from the stellar position, though the Keplerian velocity field will remain centered about the star.
In conclusion, our SA analysis, in addition to spectroscopic modeling and protoplanetary disk chemical modeling, suggest a scenario in which we observed hot water emission from the disk atmosphere above the inner, terrestrial planet region of a protoplanetary disk. Future work will include modeling additional factors such as a disk wind, fully exploring the parameter space, and exploring constraints on OH emission which was detected but for which we do not detect an SA signal.
L.B., E.L.G., and M.T. gratefully acknowledge support from NSF's Stellar Astronomy program and the American Recovery and Reinvestment Act of 2009 (NSF 0908230), JPL (RSA No: 1423736), and the NASA Exobiology and Evolutionary Biology program (NNX07AK38G). We also acknowledge the contributions made by undergraduate research assistants Bryant Dentinger and Nick Kraftor. Data presented herein were obtained at the W. M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W. M. Keck Foundation. The authors recognize and acknowledge the very significant cultural role and reverence that the summit of Mauna Kea has always had within the indigenous Hawaiian community.