ABSTRACT
T Tauri stars are low mass young stars that may serve as analogs to the early solar system. Observations of organic molecules in the protoplanetary disks surrounding T Tauri stars are important for characterizing the chemical and physical processes that lead to planet formation. Searches for undetected molecules, particularly in the inner, planet forming regions of these disks are important for testing protoplanetary disk chemical models and for understanding the evolution of volatiles through the star and planet formation process. We used NIRSPEC on Keck 2 to perform a high resolution (λ/Δλ ∼ 25,000) L-band survey of T Tauri star GV Tau N. This object is one of two in which the simple organic molecules HCN and C2H2 have been reported in absorption in the warm molecular layer of the protoplanetary disk. In this Letter, we report the first detection of methane, CH4, in a protoplanetary disk. Specifically, we detected the ν3 band in absorption. We determined a rotational temperature of 750 ± 50 K and column density of (2.8 ± 0.2) × 1017 cm−2. Our results imply that CH4 originates in the warm molecular layer of the inner protoplanetary disk.
Export citation and abstract BibTeX RIS
1. INTRODUCTION
T Tauri stars are low-mass, young stars that are often surrounded by protoplanetary disks. They may serve as analogs to the early solar system and their chemical composition, especially in the inner planet-forming regions, is of particular importance in understanding the role disks play in the composition of planets, comets, and asteroids. While CO gas has been studied in disks for many years, only recently have water and simple organic molecules like HCN and C2H2 been detected in the inner regions of T Tauri disks using the Spitzer Space Telescope (e.g., Najita et al. 2013; Carr & Najita 2011, 2008; Salyk et al. 2011; Pontoppidan et al. 2010) and ground-based, high-resolution, near-infrared spectroscopy (Lahuis et al. 2006; Gibb et al. 2007; Mandell et al. 2012). Only upper limits have been reported for many other molecules that are predicted to be abundant in disks (Bast et al. 2013).
GV Tau is a T Tauri binary system in the L1524 molecular cloud at a distance of ∼140 pc. The northern component, GV Tau N, has been of particular interest in recent years. It is one of only two protoplanetary disks to date for which organic molecules (HCN and C2H2) have been reported in the gas phase in absorption (Gibb et al. 2007; Doppmann et al. 2008). Objects must have a nearly edge-on inclination for the disk to be sampled via absorption line spectroscopy. Multi-wavelength, high-resolution observations of GV Tau support the conclusion that the northern component must have an orientation that is very close to 90° (Roccatagliata et al. 2011). The nearly edge-on orientation is rare, but offers better opportunities to detect molecules since the bright near-infrared continuum in T Tauri stars generally leads to low line contrast for emission lines. We note that even objects with a favorable (nearly edge-on) geometry often do not exhibit molecular absorption lines for organic molecules, with DG Tau B being one example in which CO and CO2 were detected in absorption, but not organic molecules (Kruger et al. 2011). This implies that the detection of organic molecules in absorption is strongly dependent on viewing angle. This makes objects like IRS 46 (Lahuis et al. 2006) and GV Tau N even more important as they are currently the only objects that can be used to test chemical model predictions of the warm molecular layer of protoplanetary disks.
We report the first detection of methane, CH4, in a protoplanetary disk. Methane has been studied in interstellar ices (see for example Boogert et al. 1996; Gibb et al. 2004; Öberg et al. 2008), in both ice and gas phases toward high mass young stellar objects (van Dishoeck et al. 1998; Boogert et al. 2004), and comets (Mumma & Charnley 2011 and references therein) and is considered to be an important molecule in astrobiology. It is a symmetric tetrahedral molecule and therefore does not have a dipole moment, so it has no pure rotational lines that can be detected using radio telescopes. It must be studied in the near-infrared via ro-vibrational transitions. The ν3 band near 3.3 μm is the strongest near-IR band and was targeted in this study.
2. OBSERVATIONS AND DATA REDUCTION
We observed GV Tau N on 22 February 2010 with the high-resolution, near-infrared cryogenic echelle spectrometer NIRSPEC at the 10 m W. M. Keck Observatory on Mauna Kea, Hawaii (McLean et al. 1998) as part of an L-band spectral survey. We used the 3 pixel (0
43) wide slit, which provides a resolving power (λ/Δλ) of ∼25,000. We used the 24'' long slit with a 12'' ABBA nod pattern. Our reduction procedures are discussed in detail in DiSanti et al. (2001), Brittain et al. (2003), and Bonev (2005). The data were flat fielded and dark subtracted, cleaned of systematically hot and dead pixels and cosmic ray hits, and resampled to align the spectral and spatial dimensions along rows and columns, respectively. The atmospheric transmittance function was modeled using the Line-By-Line Radiative Transfer Model (Clough et al. 2005), which is optimized for Mauna Kea's atmospheric conditions (Villanueva et al. 2011). It was scaled to the observed continuum (generally defined as a second order baseline through regions devoid of spectral features or telluric lines) and used to determine the column burdens of atmospheric species and to perform wavelength calibrations. Figures 1 and 2 show the normalized residual ((spectrum–telluric model)/continuum), still convolved with the telluric transmittance.
Figure 1. Spectral panels showing several transitions of CH4 in GV Tau N. The normalized (flux/continuum) residual spectrum (after subtracting the telluric model) is shown in black in each panel. Data points with transmittance less than 50% have been removed. Overplotted is the best-fit synthetic model for CH4 (red). Synthetic models for water (orange) and HCN (blue), and the photospheric model for GV Tau N (purple, offset vertically (+0.05) for clarity, see Section 2) are also shown, illustrating that most absorption features are accounted for. The telluric model used to derive the residual is shown in green, scaled (*0.04) vertically for clarity.
Download figure:
Standard image High-resolution image
Figure 2. Spectral panels showing several additional transitions of CH4 in GV Tau N. The normalized (flux/continuum) residual spectrum (after subtracting the telluric model) is shown in black in each panel. Data points with transmittance less than 50% have been removed. Overplotted is the best-fit synthetic model for CH4 (red). Synthetic models for water (orange) and HCN (blue), and the photospheric model for GV Tau N (purple, offset vertically (+0.05) for clarity, see Section 2) are also shown, illustrating that most absorption features are accounted for. The telluric model used to derive the residual is shown in green, scaled (*0.04) vertically for clarity.
Download figure:
Standard image High-resolution image
Late spectral type stars like GV Tau N can potentially exhibit significant photospheric absorption lines in the near infrared, which must be removed or flagged prior to analysis of the circumstellar features. To check for photospheric absorption lines, we adopted the method of Doppmann et al. (2005, 2008) and Najita et al. (2008), generating a synthetic stellar photosphere spectrum as described by Horne et al. (2012) using the MOOG stellar synthesis program (Sneden 1973) in conjunction with NEXTGEN stellar atmospheric models (Hauschildt et al. 1999) referencing the Kurucz atomic line lists (Kurucz 1979, 1993). Complementary models were created using the SPECTRUM program (Gray & Corbally 1994) and were used to confirm our results. We used a K3 stellar model, consistent with the estimated spectral type for GV Tau N (Goodrich 1986), with Teff = 3600 K, log g = 4.0, vsin i = 15 km s−1, and veiling = 10.5, similar to the values found by Doppmann et al. (2008). The synthetic stellar model was then used to determine which transitions were blended with strong stellar features so we could eliminate those from the analysis. In general, due to the high veiling toward GV Tau N (veiling = 10.5), most photospheric absorption lines are relatively weak. It is shown as the purple line in Figures 1 and 2, offset vertically for clarity.
3. ANALYSIS
We detected many lines (up to J = 20) of the ν3 band of CH4 in absorption. The line profiles are consistent with the gas being unresolved at a single Doppler shifted position, which is what one would expect from absorption of material in Keplerian rotation where the primary velocity component is transverse to the line-of-sight. Lines that were at good transmittance (>50%) and were not significantly blended with other species (such as water and HCN) or with photospheric lines could be well modeled and were included in the analysis. The high Doppler shift of GV Tau N at the time of the observations (∼55 km s−1) shifted many transitions well out of their telluric counterparts. Most of the lines in this analysis are from the R-branch, which are generally stronger than the P-branch lines (with higher Einstein–As). Each line (excepting the R0, R1, and P1 lines) is comprised of multiple transitions. For example, the R11 line has 21 transitions contributing to the equivalent width, though only 11 dominate. The R-branch transitions are generally close together, blending at the resolution of NIRSPEC, while the P-branch lines have a greater frequency spread. Since the telluric P-branch lines of methane are also spread in frequency, this results in the additional complicating factor that many P-branch absorptions are Doppler shifted into neighboring telluric lines. This can be seen in the bottom three panels of Figure 2, which show the normalized residual (telluric subtracted) spectra in the regions of selected CH4 lines (black) with the telluric model that was subtracted shown in green and scaled vertically for clarity.
To analyze the spectra, we generated synthetic local thermodynamic equilibrium (LTE) CH4 models for a range of temperatures. The models were Doppler shifted to the velocity of GV Tau N, scaled by the telluric transmittance, convolved to the resolution of the data, and scaled to fit the residual until the best-fit model was found that minimized the χ2 value. The molecular parameters, including line positions, Einstein-As, and intrinsic line strengths are from the 2012 update to the HITRAN molecular database (Rothman et al. 2009). We determine that all CH4 transitions are optically thin (τ < 1) for reasonable values of the intrinsic line widths (all b > 0.6 km s−1). Doppmann et al. (2008) found that the HCN spectrum for GV Tau N was well fit with a microturbulent line broadening of 3 km s−1, supporting our conclusion of optically thin CH4 absorption. Figures 1 and 2 show the residual spectra in the regions of selected CH4 lines. Overplotted in red is the best-fit (750 K) synthetic CH4 model. Also overplotted are the transmittance model (green), scaled (×0.04) for clarity, and the Kurucz photospheric model (purple), offset vertically (+0.05) for clarity. LTE water (orange) and HCN (blue) models are also overplotted, illustrating that most absorption features are accounted for with a small number of molecules. The full analysis of water and other molecules will be presented in a forthcoming paper (E. L. Gibb et al., in preparation). They are shown here for illustrative purposes and to indicate where CH4 is blended with other species.
4. DISCUSSION
We present the first detection of methane, CH4, in a protoplanetary disk. We detected ν3 R branch lines up to J'' = 20. The P-branch lines are generally weaker and more spread out in frequency than the R-branch lines and we detected fewer of them with confidence. We derive a best-fit rotational temperature of 750 ± 50 K and a total column density N(CH4) = (2.8 ± 0.2) × 1017 cm−2 (1σ). We note that Gibb et al. (2007) reported a CH4 upper limit of 2.2 × 1016 cm−2 by sampling the low-J (R0, R1, and R2) lines and assuming a low (115 K) rotational temperature. The higher rotational temperature of the gas resolves the inconsistency in derived column densities.
4.1. Comparison to Chemical Models
Is a CH4 detection consistent with what is expected from chemical models of disks around low mass young stars? Several models of chemistry in protoplanetary disks predicting high abundances of methane have been published in recent years (Agúndez et al. 2008; Willacy & Woods 2009; Heinzeller et al. 2011). The models differ in their details but generally include high energy irradiation of the disk by a combination of UV and X-ray radiation that contribute to driving the gas phase chemistry. While Agúndez et al. (2008) ran separate models for FUV and X-ray illuminated disks to investigate their affects on the chemistry, most authors, including Willacy & Woods (2009) and Heinzeller et al. (2011), included both simultaneously. Only the most recent models have started to include radial viscous accretion, vertical turbulent mixing, and vertical disk winds to study the affects of bulk gas motions on disk chemical structure (Heinzeller et al. 2011). Physically, regardless of the details of the models, three chemically distinct regions are predicted. The midplane of the disk is where volatiles are expected to freeze onto dust grains beyond the snow line. The disk atmosphere, exposed to FUV and X-ray radiation, gives rise to a photon-dominated region characterized by ionization and dissociation products. Between those two layers is the warm molecular layer, which is somewhat protected from ionizing radiation by dust and polycyclic aromatic hydrocarbons in the surface region. It is in this warm molecular layer that organic molecules are predicted to be abundant. The reader is referred to these papers for additional details.
CH4 is produced in the gas phase at high temperatures, above ∼700 K, by a reaction sequence that begins with the formation of C from the dissociation of CO followed by a series of reactions with H2, leading to CH4. CH4 may also be formed efficiently on dust grains at the low temperatures found near the midplane (Öberg et al. 2008), and diffusion can act to increase the fractional abundance at greater heights above the midplane (Heinzeller et al. 2011). Near the disk surface, UV and X-ray photons destroy CH4 on short timescales. It is therefore expected that CH4 will predominantly be found in the gas phase in the warm molecular layer at scale heights less than ∼0.3 AU in the inner protoplanetary disk (see Figure 7 of Heinzeller et al. 2011), consistent with interpretation of the locations of other organic molecules (HCN and C2H2) and with the observation that GV Tau N is very close to edge-on (Roccatagliata et al. 2011).
Agúndez et al. (2008) predicted a high abundance of CH4 in their X-ray illuminated disk models for kinetic temperatures similar to those derived for CH4, HCN, and C2H2 (though the FUV illuminated disk models predict much lower CH4 abundances at later times, see their Figure 1). Indeed, for some models the abundances of HCN, C2H2, and CH4 are comparable. Heinzeller et al. (2011) report vertical column densities for several species in their models at a radius of 1.3 AU. CH4/HCN varies from 0.06 for their accretion model to 1.5 for the model that includes turbulent mixing. Willacy & Woods (2009) on the other hand, predicted CH4/HCN ∼ 0.007–0.06 at 1 AU. The low ratio is primarily due to a very high vertical column density predicted for HCN (1.1 × 1020 cm−2 as compared to 4.3 × 1019 cm−2 in the Heinzeller et al. 2011 model). At larger radial distances, they predict that HCN freezes out, giving a much lower gas phase column density and a correspondingly high CH4/HCN ratio (∼350 at 5 AU, for example). Regardless, a common theme in these models is a high abundance of CH4 in the inner region of the disk in the warm molecular layer above the midplane where temperatures, though model dependent, exceed several hundred Kelvin.
We emphasize that these chemical disk models report vertical column densities while our absorption line study measures column densities through the disk at high inclination angles. Our derived abundances are expected to depend sensitively on the height above the midplane sampled by the observations. This can be seen, for example, in Figure 3 of Walsh et al. (2010), which shows relative abundances of several molecules (not including CH4) as a function of height above the midplane and radial distance from the star. Clearly, additional modeling is needed for meaningful interpretation of absorption line studies in disks around young stars. Specifically, temperature profiles and column densities as a function of inclination for nearly edge-on sources are vital to fully interpret our results and to test models. Such models may also help address the non-detections of organic molecules in other nearly edge-on sources like DG Tau B. Nevertheless, given the limitations of currently available models, our CH4 results are consistent with the expected high abundance of warm CH4 gas that implies an origin in the warm molecular layer of the inner disk.
4.2. Comparison to Other Observations
Are our observations consistent with other measurements toward GV Tau N? Bast et al. (2013) analyzed Spitzer detections of HCN, C2H2, and CO2 absorption in GV Tau N and found rotational temperatures of about 440 K, 720 K, and 250 K, respectively. The temperature differences are likely due to different radial distributions for each molecule. The lower temperature for CO2, for example, may be due to the column density peaking at a larger distance from the star than HCN and C2H2, which is consistent with the model predictions of Agúndez et al. (2008) and Heinzeller et al. (2011) (see Section 4.1). Gibb et al. (2007) reported Trot ∼ 260 K for 13CO, which is also consistent with the column density peaking at a greater radial distance as expected from chemical models. Doppmann et al. (2008) reported that their L-band data were well fit with Trot = 550 K for HCN. Our 750 K CH4 rotational temperature is somewhat warmer than HCN, but does imply a similar origin in the inner disk region. Agúndez et al. (2008) found that CH4 vertical column densities peaked within ∼0.5 AU of the disk while the HCN vertical column density remained high (∼1016 cm−2) beyond 1 AU, consistent with the somewhat warmer rotational temperature derived for CH4.
Since absorption line studies sample a long column of material, and since abundance distributions in disks have significant radial and vertical dependences, abundance ratios calculated by taking ratios of column densities of different species that have different rotational temperatures (and are therefore not likely to be cospatial) may not be meaningful. However, since both HCN and C2H2 sample material reasonably similar in temperature to CH4, and since chemical models predict somewhat similar spatial distributions for these molecules in the inner regions of protoplanetary disks, an abundance comparison may be of interest. Bast et al. (2013) reported N(C2H2) = (1.4 ± 0.3) × 1016 cm−2 and N(HCN) = (1.8 ± 0.4) × 1016 cm−2 for GV Tau based on mid-infrared Spitzer Space Telescope data. Doppmann et al. (2008) reported N(HCN) = 1.5 × 1017 cm−2 based on ground-based high-resolution near-infrared spectroscopy. Gibb et al. (2007) reported N(HCN) = 3.7 × 1016 cm−2 based on an analysis of low-J lines, but when the higher J lines are sampled, the resulting rotational temperature and column density are more consistent with those reported by Doppmann et al. (2008) (these results will be published in a forthcoming paper, E. L. Gibb et al., in preparation). If we assume that the organic molecules originate along the same absorbing column and compare only the ground-based near-IR column densities, we find HCN/CH4 ∼ 50%, reasonably consistent with the X-ray illuminated, warm gas temperature models of Agúndez et al. (2008) (see their Figure 1) and the turbulent mixing model of Heinzeller et al. (2011), though as discussed above more modeling work is needed before meaningful comparisons can be made between our observations and chemical models.
4.3. Comparison to Comets
CH4, HCN, and C2H2 have also been studied in comets. While abundances of organics vary from comet to comet, HCN/CH4 is generally found to vary between 10% and 50% and C2H2/CH4 is about 7%–35% (Gibb et al. 2012; Mumma & Charnley 2011 and references therein), similar to the value found for the disk of GV Tau N. Comets are thought to represent the final midplane abundances of molecules, but the presence of crystalline silicates strongly supports the hypothesis that radial and vertical transport mechanisms were important to the overall final compositions (see Mumma & Charnley 2011 and references therein). Hence, abundances of HCN, C2H2, and CH4 in comets likely represent a combination of ice processing in the disk midplane and gas processing in the warm molecular layer and a comparison to disks is justified.
5. SUMMARY AND CONCLUSIONS
We report the first detection of CH4 via the ν3 band in a protoplanetary disk using high-resolution, near-infrared L-band spectroscopy. The rotational temperature (750 ± 50 K) is similar to rotational temperatures found for other molecules that have been observed in the inner regions of disks around T Tauri stars and is consistent with a location in the warm molecular layer of the inner region of the disk, though the slightly higher temperature for CH4 implies a peak abundance closer to the star than for HCN, CO2, and CO.
Abundance comparisons indicate similarities to cometary ices, which may suggest thermal and energetic processing of cometary material, though further observations of both disks and comets are needed to test such a scenario. The current study also emphasizes the need for chemical models that report column densities along high inclinations or abundances as a function of vertical and radial position in the disk.
The 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. E.L.G. was supported by NSF Astronomy grant AST-0908230 and NASA Exobiology grant NNX07AK38G (PI: D. C. B. Whittet). D.H. was supported by NSF Astronomy grant AST 05-07419.