A Search for Spatially Resolved Infrared Rovibrational Molecular Hydrogen Emission from the Disks of Young Stars

The bulk gas constituent in circumstellar disks around young Sun-like stars is in the form of molecular hydrogen. Table 1 presents the excitation and origins of the UV, near-IR, and mid-IR molecular hydrogen features from the environments of CTTSs. This table serves as a summary of key H₂ features that have been detected or interpreted in CTTSs; it is not meant to be a complete listing of all theorized excitation mechanisms for H₂ from young stars. Figure 1(a) shows a schematic cross-cut of an edge-on disk system, highlighting the location of UV, near-IR, and mid-IR H₂ emission from the inner ∼100 au of a young single-star system. The structure in this figure is based on the density, temperature, and chemical models of circumstellar disks developed in the past two decades by Nomura & Millar (2005), Nomura et al. (2007), Dullemond et al. (2007), Woitke et al. (2009), Walsh et al. (2010), and Woitke et al. (2018). The UV H₂ traces hot low-density gas predominantly excited by Lyα pumping in the inner 10 au disk regions and the inner outflows. The UV H₂ can also arise from nonthermal Lyα pumping in regions of the collimated jets and inner winds where the H₂ is not dissociated by shocks or high temperatures. In regions with significant UV H₂ emission the gas temperatures are typically in the 2000–3000 K range, but nonthermal pumping of the H₂ by Lyα is the dominant excitation mechanism over collisional excitation. Near-IR H₂ measures denser warm gas in LTE in the upper disk layers excited by central stellar flux, and in the shock-excited inner outflows. The mid-IR H₂ traces cooler, denser gas closer to the disk midplane within ∼30 au of the central star. Figure 1(b) is a representation of the H₂ emission from an equal-mass 100 au separation binary star. The disk emissions shown in Figure 1(a) are also present here, though each circumstellar disk is dynamically truncated at ∼30 au (Artymowicz & Lubow 1994). The system is encompassed by a circumbinary ring with 300 au radius, and the inner circumstellar disks are fed by accretion streamers that funnel material into the inner regions (Artymowicz & Lubow 1994, 1996). Shocked rovibrational H₂ emission from material in accretion infall is also postulated as a viable excitation mechanism in dynamically complex young star multiples.

Zoom In Zoom Out Reset image size

At UV wavelengths, the electronic transitions of H₂ are fluorescently pumped by the strong Lα emission from young stars and trace very warm (>2000 K) low-density (<10³) gas in the central disk. In France et al. (2012), UV H₂ emission was measured in 100% of the actively accreting stars surveyed with the HST COS. They found a correlation between H₂ flux and stellar Lyα and accretion-generated CIV luminosity, as well as a decreasing trend of UV H₂ emission with age from ∼1 to 10 Myr. Analysis of the UV H₂ emission profiles reveals an origin within the inner ∼3 au disk from interaction between the strong Lyα emission with the molecular disk surface. Only three stars overlap between our present study and the France et al. (2012) sample: AA Tau, GM Aur, and LkCa 15. Conversely, the pure rotational H₂ transitions in the mid-IR trace cooler, dense molecular hydrogen at deeper depths near the circumstellar disk midplane. The v = 0–0 S(2), S(1), and S(0) lines at 12.28, 17.03, and 28.0 μm are used to quantify the molecular hydrogen component in dense extended disks. The Spitzer Space Telescope IRS instrument with the R ∼ 600 high-resolution mode was not quite sensitive enough to detect these very weak quadrupolar rotational H₂ transitions above bright mid-IR continuum emission in most of the systems studied in the CTTS surveys (Baldovin-Saavedra et al. 2011). From our present survey sample, only AA Tau, UY Aur, and AB Aur have measured rotational H₂ transitions from Spitzer or high spectral resolution ground-based measurements (Bitner et al. 2007; Carr & Najita 2011). In AB Aur, the measured mid-IR rotational transitions indicate a gas temperature of 670 K and an H₂ emission location within 18 au in the circumstellar disk (Bitner et al. 2007). Woitke et al. (2018) found that strong rotational H₂ emission requires very large column densities and a large temperature contrast between the gas and dust at deeper layers inside disks, which they do not see in their models.

The near-IR rovibrational H₂ transitions provide an important cooling mechanism in the inner <3 au terrestrial regions of circumstellar disks (Woitke et al. 2009). While the rovibrational H₂ emission has been studied in the spectra of young stars for decades, clear identification of the spatially resolved H₂ arising from circumstellar disks has proven difficult. In many of the youngest systems, these near-IR H₂ features are stronger or more readily detected from shock-excited emission associated with or encompassing the outflows from the stars (Beck et al. 2008). Additionally, disk-excited rovibrational H₂ emission arises predominantly from the low- to moderate-density heated disk surface layer (10³–10⁵ molecules cm⁻³; Figure 1(a)), and it is not expected to extend beyond ∼50 au from the central star (Nomura & Millar 2005).

On UT 2001 November 15 and 16 high-resolution spectroscopy was acquired for using the CGS4 infrared spectrograph in echelle mode at the United Kingdom Infrared Telescope (UKIRT) on Maunakea, Hawaii (Table 1). Using that 1-pixel-wide slit (∼061), long-slit observations were made using a standard ABBA nod pattern for the efficient removal of sky emission features, dark current, and sky background. The velocity sampling is 3.7 km s⁻¹ per pixel in the wavelength-calibrated data, yielding an instrumental 2-pixel resolving power of R = 40,000. The data were reduced using the standard packages in STARLINK software package CGS4DR and Figaro. These data were significantly affected by fringing, most of which was successfully removed during the reduction process, which necessitated a top-hat data smoothing to an effective resolving power of R = 18,000 and a 2-pixel velocity resolution of v = 35 km s⁻¹.

On UT 2010 September 25 a follow-up high-resolution spectrum of one target, GM Aur, was obtained with the CSHELL spectrograph at the Infrared Telescope Facility on Maunakea, Hawaii (Table 1). Using the 5-pixel-wide, 10 slit, we observed the 2.12 μm emission from GM Aur in a total of 42 minutes on source. These data were reduced with a CSHELL specific IDL package called CSHELLEXT (C. Bender 2019, private communication). Alternatively, the data were also reduced with standard IRAF routines. A comparison of the two procedures found them to be similar within the uncertainties of the processed data. The instrumental spectral sampling gave a 2-pixel resolving power of R = 21,000 for this observation, or v = 28 km s⁻¹. The spectra presented here are the product of the IDL package.

All of the H₂ emission-line velocities are corrected for heliocentric motion at the time of the observation and placed into the stellar rest frame by removing the radial velocity of the star (Nguyen et al. 2012; Hartmann et al. 1986). All velocities presented in this project are v_Helio rather than v_LSR.

Integral field spectroscopic observations of nine CTTSs were obtained using the NIFS at the Gemini North Frederick C. Gillette Telescope on Maunakea, Hawaii (Table 1). NIFS is an image-slicing IFU fed by Gemini's near-IR adaptive optics system, Altair, which is used to obtain integral field spectroscopy at spatial resolutions of ≤01 with a 2-pixel spectral resolving power of R ∼ 5300 at 2.2 μm (McGregor et al. 2003). The NIFS field is 3'' × 3'' in size, and the individual IFU pixels are 01 × 004 on the sky. Data were obtained at the standard K-band wavelength setting for a spectral range of 2.010–2.455 μm. Each of the stars observed for this program was bright enough to serve as their own wave front reference stars for the adaptive optics, and observations were typically acquired in natural seeing better than ∼085 for excellent AO correction.

The NIFS data for this project was acquired during three observing terms at Gemini under program IDs GN-2007B-Q-40, GN-2009A-Q-100, and GN-2009B-Q-40. All nine sources observed possess detectable levels of v = 1–0 S(1) rovibrational H₂ line emission in long-slit high-resolution data such as those presented here and in our previous work (AA Tau, GG Tau A, GM Aur, UY Aur, DoAr 21, GG Tau A, LkCa 15) or reported in the literature (LkHα 264). One source, AB Aur, had a published detection of mid-IR rotational emission (Bitner et al. 2008). The IFU data obtained are summarized in Table 1. Each star was observed with many short exposures to avoid saturation on the bright central source. In Columns (3) and (4) of Table 2 the exposure times and number of individual exposures are listed. Column (5) presents the final on-source median exposure time for each star. All NIFS data on the young stars were acquired by offsetting to a blank-sky field to measure background flux brightness every roughly third image, or approximately every 2–3 minutes. For each observation, a standard set of calibrations was acquired using the Gemini facility calibration unit, GCAL. The data for most stars were obtained over several nights, and associated supporting calibration data were acquired for each night. Column (6) of Table 1 presents the adopted K-band magnitude for each star. In most cases, the K-band magnitudes are adopted from data acquired with the WIYN High-Resolution Infrared Camera (WHIRC), and Column (7) presents the source for these measurements. The supporting photometry was acquired in 2009 September with the 3.5 m WIYN telescope and the WHIRC using the K' filter and the UKIRT faint standard FS 18 for flux calibration.

Table 2.  Log of Observations

Star	Obs. Date	Exp. Time(s)/	No. Exp.	Total On-source	K-band	Magnitude
Name		No. Co-adds		Exp. Time (s)	Magnitude	Source
			UKIRT CGS4
AA Tau	2001 Nov 15	30.0	16	480	⋯	⋯
GG Tau A	2001 Nov 15	15.0	16	240	⋯	⋯
GM Aur	2001 Nov 16	30.0	8	240	⋯	⋯
UY Aur	2001 Nov 16	30.0	9	270	⋯	⋯
			IRTF CSHELL
GM Aur	2010 Sep 25	180.0	14	2520.0	⋯	⋯
			Gemini NIFS
AA Tau	2007 Oct 14, Nov 11,	40.0/1	31	1240	7.9	WIYN+WHIRC
	2008 Jan 4, 5
AB Aura	2009 Nov 9, 19	40.0/1	41	1640	4.23	2MASS
	2009 Dec 13, 26
DoAr 21	2009 Sep 4, 5	40.0/1	76	3040	6.23	2MASS
GG Tau A	2009 Aug 14, Dec 1	40.0/1	78	3120	7.7b	WIYN+WHIRC
GM Aur	2009 Oct 18, 20, 24	40.0/1	123	4920	8.5	WIYN+WHIRC
LkCa 15	2007 Nov 12, 15, 24	40.0/1	54	2160	8.2	2MASS
LkHα 264	2009 Aug 10, 11	40.0/1	105	4200	9.05	WIYN+WHIRC
UY Aur	2009 Aug 22	40.0/1	104	4160	7.6b	WIYN+WHIRC
V773 Tau	2009 Aug 11, 13, 22	20.0/2	119	4760	6.7b	WIYN+WHIRC

Notes.

^aAB Aur was observed using the 02-diameter occulting disk in the NIFS to attenuate the continuum flux from the bright star. ^bThe presented K-band magnitude values for GG Tau A, UY Aur, and V773 Tau include flux contributions from all stars in these multiple-star systems.

Download table as:  ASCIITypeset image

The raw NIFS IFU frames were reduced using the NIFS tasks in the Gemini IRAF package.³ The basic reduction steps for processing NIFS raw data into flat-fielded, sky-subtracted, rectified pixels in a 3D data cube are described in detail in Beck et al. (2008). We refer to this manuscript for description of the main processing steps, nfreduce, nffixbad, nffitcoords, and nftransform, which execute the basic NIFS IFU reductions. For each of the science targets, we observed an A0 spectral type star with NIFS for removal of telluric absorption features. These data were also processed using the above tasks in the Gemini IRAF package. The A0-type stars were chosen based on their spectral types defined in the Hipparcos catalog and because they provided a good match to the airmass of the science target. A pseudo-long-slit spectrum of each calibration star was extracted from the data using a 10-radius circular aperture using the IRAF task nfextract. The spectral continuum shape and atomic H i absorption features were removed using task nffixa0, which was developed for this purpose. The science data cubes of the young stellar objects (YSOs) were divided by the corrected spectra of the A0 calibrators (in the wavelength dimension) using the task nftelluric.

After the data were corrected for telluric absorption, they were built into individual data cubes using the nifcube task with an 004 square pixel spatial sampling. The data for each source were collected using a small dither pattern to allow for the effective removal of hot/cold pixels, to even out the pixel-to-pixel sensitivity variations, and to improve the overall pixel sampling in the spatial dimensions. The resulting data cubes were registered and co-added using a 3D shift-and-add routine, which determines the stellar central position in each data cube; shifts the spatial axes to a common, central location; and then median-combines all the spectral data in λ-dimension corresponding to each spaxel. For optimal signal-to-noise ratio (S/N) on faint extended H₂ emission, individual NIFS data frames that were observed with instantaneous measured natural seeing of worse than 085 were excluded from the final co-added data cubes. As a result, the processed data had better spatial resolution (∼01) and rounder point-spread function (PSF) shapes than if all of the raw NIFS data had been included. The final data products demonstrate this observing strategy and data reduction method to be particularly powerful, providing high-contrast images that permit detection of low-level line emission in the environments of bright young stars.

After the cubes were combined into a single 3D product, they were flux-calibrated as a last step in the reduction. This was accomplished by extracting a 1D spectrum over the full spatial field, convolving this 1D spectrum with a K-band filter function (filter function depends on the source of photometry; see Column (7) in Table 2), and deriving and applying a relative flux calibration scaling using the adopted K-band magnitude (Column (6) of Table 2) and the instrument flux sensitivity function. Whenever possible, we used our own photometry taken nearby in time to calibrate the fluxes of these young stars, since YSOs are known to be variable in infrared flux. Typical uncertainties in our observed K-band magnitudes are at the 0.1 mag level. For the sources for which we were not able to acquire nearly simultaneous photometry, we used magnitudes from the Two Micron All Sky Survey (2MASS; Skrutskie et al. 2006). We conservatively estimate the uncertainties in the flux calibrations to be ∼5%–10% when using the WHIRC photometry and ∼15% for sources where we relied on 2MASS magnitudes—AB Aur, GM Aur, and DoAr 21.

AB Aur was observed with the NIFS 02 occulting disk in the entrance pupil wheel to block the flux from the bright central star in this system. The continuum emission from AB Aur would have saturated in the minimum available exposure time if the occulting disk had not been used. The occulting disk is a dark obscuring spot on a transmissive substrate. The light that passes through the substrate creates a fringing-like modulation that imprints a sinusoidal pattern on the continuum flux data at the ∼3%–4% level. This sinusoidal modulation of the stellar flux can be mitigated by acquiring an occulting disk flat, which can lessen this continuum structure to the ∼0.5% level. It is also important to note that the occulting disk is slightly transmissive, attenuating the central flux by a factor of 10⁴. The central flux was recovered by correcting the AB Aur data using a flat-field exposure taken with the occulting disk in the beam simultaneously with the observed data. Given the attenuation and fringing associated with the occulting disk, the value of central flux measured for AB Aur is very uncertain. The approximate absolute flux calibration for AB Aur was performed using the target acquisition images acquired with a neutral density filter. The flux calibration sensitivity of the AB Aur data is correspondingly less certain than the other sources, and all line fluxes determined for AB Aur are lower limits owing to the attenuation of the central H₂ emission.

By inspecting the wavelength calibration and accuracy of sky emission-line positions in raw exposures, we estimate that the absolute velocity accuracy of the individual IFU cubes is ∼9–12 km s⁻¹, or roughly one-third of a spectral pixel for a given night. However, over multiple nights, the repositioning accuracy of the NIFS IFU grating wheel is repeatable only to ∼±0.75 pixels. Therefore, data acquired for the same source, but on different nights, were not shifted and co-added in the wavelength dimension because of very weak signal from H₂ emission and to preserve velocity resolution. It is important to note that the accuracy of the velocity centroids of the emission lines in the co-added multinight data is clearly affected by the grating repositioning and the resulting shifts in the wavelength dimension. However, all of the NIFS data on UY Aur were acquired during the same night with no movement of the grating wheel. As a result, the UY Aur data have the single-night kinematic absolute accuracy of 9–12 km s⁻¹ and relative in-field kinematic resolution near this level. For all other sources, the estimated velocity accuracy of our IFU data is at the level of 1.0–1.2 pixels, or 28–35 km s⁻¹. In these cases, the relative kinematic shifts within the spatial IFU field are likely accurate to approximately half of this value, or ∼14–17 km s⁻¹. In the following sections, aside from UY Aur, only emission-line velocity shifts of greater than the absolute accuracy level are taken to be genuine. Additionally, all of the discussed H₂ emission-line velocities are corrected for heliocentric motion at the time of the observation and placed into the stellar rest frame by removing the radial velocity of the star (Hartmann et al. 1986; Nguyen et al. 2012). All velocities presented in this project are v_Helio rather than v_LSR.

While investigating the IFU data cubes for this project, it became apparent that we discovered and characterized a previously unknown optical ghost in the NIFS instrument. This ghost is seen in all of the K-band IFU spectral imaging cubes of our target CTTSs. This ghost is presented and described in more detail in the Appendix.

Figure 2 presents the H₂ emission-line detections in AA Tau, GG Tau A, GM Aur, and UY Aur from the high-resolution spectroscopy. Table 3 summarizes the velocity properties of the H₂ line emission. Line profile velocity centroids and velocity widths were derived by fitting Gaussians to the observed features. To place a measure on uncertainties in this process, the fits were carried out using three different tools and slightly different parameters for continuum level identification. This investigation showed that the measured velocity values are accurate to about 3 km s⁻¹ (0.2 pixels) for both CGS4 and CSHELL. AA Tau shows a moderate blueshifted H₂ line centroid, and the other three systems are largely consistent within 1σ–2σ of the stellar rest velocity, which is at 0 km s⁻¹ in Figure 2. The line profile width for AA Tau is perhaps marginally resolved, and the other three systems are consistent with spectrally unresolved H₂ emission in our data sets. The purpose of these high-resolution measurements was to identify clear detection of the v = 1–0 S(1) rovibrational H₂ and confirm the existence of low-velocity spectrally unresolved features indicative of potentially disk-bearing molecular hydrogen gas.

Zoom In Zoom Out Reset image size

The goal of this study is to characterize the spatial extent of the H₂ emission previously detected in these systems and to isolate emission arising from the inner circumstellar disks. Figure 3 presents the 2.12 μm continuum images of the three young multiple-star systems that we spatially resolved in this study. The point-like continuum images of the single stars are not included. These images were derived by fitting the average continuum level on the blueward and redward side of the H₂ line using a second-order polynomial and integrating in wavelength across the emission feature. Our AO-fed imaging spectroscopy, coupled with the multislice dithered observing strategy, results in nicely rounded spatial PSFs typically with spatial FWHM values of 008–010. DoAr 21, GG Tau Ab, and V773 Tau have unresolved companions that were below the spatial sensitivity limit in our data (Loinard et al. 2008; Boden et al. 2012; Di Folco et al. 2014).

Zoom In Zoom Out Reset image size

Of the nine stars observed, seven have H₂ emission that was clearly spatially resolved in the IFU data. Figure 4 presents continuum-subtracted maps of the 2.12 μm flux showing the distribution of the H₂-emitting gas in AA Tau, DoAr 21, GG Tau A, GM Aur, LkHα 264, UY Aur, and V773 Tau. The H₂ images were made by first fitting the continuum level on either side of the emission line and interpolating between these points to estimate the continuum flux at the position of the feature. The continuum flux was then subtracted from the spectrum over the wavelength range covered by the line, and the remaining line flux was integrated over the entire feature. These steps were performed for the λ-dimension of each pixel in the data cube. The final 2D images presented in Figure 4 isolate the 2.12 μm line emission from the H₂ gas in seven of the systems.

Zoom In Zoom Out Reset image size

AA Tau, UY Aur, and V773 Tau all have discrete knots of H₂ emission at extended distances from the stars. The detected emission from AA Tau shows an arc extending to the east–southeast and a bright knot of emission at a position angle of ∼185°, measured east of north. The extension of emission is aligned with the known near edge-on disk encircling AA Tau and may arise from the disk surface in the inner 03 regions (<50 au) from the star. The bright knot of emission is aligned with the direction of the known microjet from the AA Tau system (Cox et al. 2013). The H₂ emission detected in DoAr 21 is entirely spatially extended, starting at angular distances of greater than 05 (∼65 au) from the central star. The H₂ exhibits a smooth, arc-like morphology and is spatially consistent with the extended dust emission discovered in resolved mid-IR maps of the system (Jensen et al. 2009). The H₂ emission in GG Tau A shows a bright arc to the northeast of the stars, with arc-like streamers that extend away from the central system. An analysis of the observed H₂ emission from the interesting GG Tau A system has been published in a previous article (Beck et al. 2012; Dutrey et al. 2016).

UY Aur exhibits the strongest integrated H₂ flux of all the stars studied in this survey (Table 4). The flux shows an extended distribution of H₂ gas entirely encompassing the stars in the binary with several extended knots of discrete emission. There is also a bright arc of emission to the south of UY Aur A that extends to more than 07 (100 au) from the star toward the east, in the direction opposite from the companion, UY Aur B. Several other arcs also extend to the south from UY Aur A and B. UY Aur is also known to drive an outflow as suggested by prior investigations (Hirth et al. 1997; Pyo et al. 2014). The H₂ emission from the V773 Tau multiple also shows an extended distribution of flux to more than 140 au from the brightest central star and knots of emission. The brightest H₂ emission in V773 Tau arises from the position of the 02 separation companion. We detect significant H₂ emission in the vicinity of both of the known infrared-luminous companions (IRCs) in the V773 Tau and UY Aur systems. The H₂ emission velocity structure observed in V773 Tau reveals, for the first time, a bipolar outflow in this system, which will be discussed more thoroughly in Section 3.6. Hence, an appreciable amount of the extended H₂ emission we observe in the AA Tau, UY Aur, and V773 Tau systems seems to be stimulated by shocks in the outflows.

Table 4.  H₂ Line Fluxes from NIFS IFU Data

Star	Extraction	1–0 S(1) Line Flux
	Aperture Radius	2.12 μm (erg cm−2 s−1)
AA Tau	124	(3.2 ± 0.1) × 10−15
AB Aura	112	>(6.6 ± 0.7) × 10−14
DoAr 21	112	(1.3 ± 0.6) × 10−15
GG Tau A	112	(6.0 ± 0.3) × 10−15
GM Aur	124	(4.9 ± 0.1) × 10−16
LkCa 15	112	<1.2 × 10−15
LkHα 264	124	(4.5 ± 0.1) × 10−15
UY Aur	092	(9.6 ± 0.4) × 10−15
V773 Tau	124	(2.6 ± 0.1) × 10−15

Note.

^aThe integrated v = 1–0 S(1) line flux for AB Aur is derived by summing the flux in a 112 aperture from the H₂ image presented in Figure 8(a).

Download table as:  ASCIITypeset image

Of the nine stars, the detected H₂ emission presented in Figure 4 for GM Aur and LkHα 264 is most similar to that expected to arise from the inner disks of T Tauri stars. The emission from these systems appears centrally compact, spatially resolved, and centered approximately on the stellar radial velocity (to our limited accuracy). The measured gas is in agreement with high-resolution detections of the 2.12 μm lines in GM Aur and LkHα 264 (Figure 2; Itoh et al. 2003; Carmona et al. 2008b). As an illustration of the extended nature of the H₂-emitting gas, we present comparisons of the encircled energy in the PSF and the PSF-subtracted noise energy to that of the continuum-subtracted H₂ in Figure 5. The PSF encircled energy is essentially a sum of the flux as a function of distance from the central star, normalized to a value of 1.0 at a 12 distance. The encircled energy for the PSF subtraction noise was estimated by executing a continuum PSF subtraction at a wavelength of 2.18 μm, in a region that is largely devoid of spectral features, summing the PSF subtraction residuals into an image, calculating the noise energy level as a function of distance from the star, and normalizing in the same manner to 1.0 at a 12 distance. The difference between the observed PSF energy and the PSF subtraction noise profiles and the H₂ profile indicates that H₂ emission is spatially resolved and extended in GM Aur and LkHα 264. The fact that the H₂ encircled energy curves for both GM Aur and LkHa 264 lie significantly below the summed PSF energy and noise curves means that the radial H₂ emission energy is spatially extended and resolved.

Zoom In Zoom Out Reset image size

Figure 6 shows the encircled energy profiles for LkCa 15, the PSF subtraction noise, and the continuum-subtracted residuals at the wavelength of the H₂ feature. The PSF subtraction noise is calculated in the same way as described for Figure 5. The PSF noise and H₂ encircled energy curves deviate from the PSF shape by at most ∼3%–4% at all distances from the star. Hence, the analysis at the H₂ wavelength shows no statistically significant difference from pure PSF subtraction residuals. We conclude that there is no appreciable H₂ emission toward LkCa 15 that is measured in the IFU data. This result is discussed in more detail in Section 4.4.6.

Zoom In Zoom Out Reset image size

Figures 7 and 8 present further investigation to detect extended near-infrared molecular hydrogen in the environment of AB Aur. The initial analysis shown in Figure 4 did not find appreciable H₂ emission greater than ∼0.5% of the integrated peak continuum flux for AB Aur. Within 01 from the star, the occulting spot used for the observations negatively affects the sensitivity to extended emission. For this analysis, we masked out this inner region. Figure 7 shows a zoomed-in view of the 2.115–2.130 μm spectral region of a continuum-subtracted small aperture with 012 × 012 spatial extent, extracted to the northwest of AB Aur at a ∼024 distance. Careful subtraction of continuum emission measured nearby in wavelength was necessary to measure the H₂ from AB Aur because the peak H₂ line emission is just 1.5% above the continuum flux measured at this single spatial position from the star. The H₂ flux measured in this small spatial aperture is included for AB Aur in Table 5. Integrated over the full IFU field of view, the very low level H₂ in AB Aur was swamped by the bright stellar flux at spatial locations that have no H₂ and by uncertainties in the continuum subtraction analysis used for Figure 4.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 5.  Peak (Narrow-aperture) H₂ Line Fluxes and 3σ Detection Limits^a

Star	Extraction	1–0 S(1) Line Flux	1–0 S(2)	1–0 S(0)	2–1 S(1)	1–0 Q(1)b	1–0 Q(2)b	1–0 Q(3)b
	Aperture Radius	2.12 μm (erg cm−2 s−1)	2.03 μm	2.22 μm	2.24 μm	2.40 μm	2.41 μm	2.42 μm
AA Tau	012	2.6 ± 0.1 × 10−16	0.30 ± 0.04	0.16 ± 0.03	0.11 ± 0.03	0.76 ± 0.18	${0.2}_{-0.12}^{+0.20}$	0.75 ± 0.15
AB Aur	012	9.5 ± 0.8 × 10−15	⋯	⋯	⋯	⋯	⋯	⋯
DoAr 21	012	2.5 ± 0.1 × 10−16	0.33 ± 0.04	0.28 ± 0.03	<0.09 (3σ)	1.1 ± 0.3	0.4 ± 0.2	0.9 ± 0.3
GG Tau A	012	4.8 ± 0.3 × 10−16	0.37 ± 0.04	0.29 ± 0.02	0.06 ± 0.02	1.1 ± 0.2	0.3 ± 0.2	1.0 ± 0.2
GM Aur	012	8.5 ± 0.4 × 10−16	<0.20	<0.24	<0.18	⋯c	⋯	⋯
LkHα 264	012	1.7 ± 0.2 × 10−16	0.27 ± 0.06	0.19 ± 0.08	<0.12	0.8 ± 0.1	0.3 ± 0.1	0.8 ± 0.2
UY Aur	012	2.7 ± 0.1 × 10−16	0.29 ± 0.03	0.28 ± 0.04	0.07 ± 0.02	0.9 ± 0.1	0.3 ± 0.1	0.8 ± 0.2
V773 Tau	012	3.4 ± 0.1 × 10−16	0.26 ± 0.03	0.26 ± 0.02	0.08 ± 0.03	⋯c	⋯	⋯

Notes.

^aThe integrated v = 1–0 S(1) line flux is presented in the third column. All other H₂ emission line fluxes are presented as ratios with respect to the measured v = 1–0 S(1) line flux. ^bQ-branch flux uncertainties are very high because of imperfect correction of narrow telluric features in the 2.40–2.44 μm region. Uncertainties in Q-branch emission fluxes were estimated by investigating the residual telluric features in the corrected spectra. This is assumed to be a lower limit on the uncertainties. ^cPoor telluric correction made it impossible to define the continuum flux level and unambiguously identify Q-branch features. Accurate flux ratio limits could not be derived.

Download table as:  ASCIITypeset image

Figure 8(a) shows the point-source continuum-subtracted and integrated emission for AB Aur at the 2.12 μm wavelength corresponding to H₂ emission. This image was created by (1) identifying two nearby 3D spatial-spatial-wavelength continuum flux regions that were modulated by the occulting disk substrate in a similar manner to the spectral region surrounding the H₂ line, (2) averaging these two regions to create a 3D continuum model, (3) scaling and subtracting this average continuum 3D data cube from the data cube centered on the H₂ emission, and (4) summing the central four wavelength channels of the continuum-subtracted H₂ data cube into a 2D image. For AB Aur, this 3D image subtraction analysis was significantly more accurate at detecting high-contrast line flux than the 1D spectral fitting method presented in Figure 4. This analysis was also more efficient at removing the slight residual structure from spectral modulation caused by using the occulting disk substrate material. Figure 8(b) shows the corresponding average uncertainty in the process used to construct Figure 8(a). Here, instead of subtracting the model data cube from the cube centered on the H₂ emission, the continuum model data cube was subtracted from another nearby continuum flux region with the same modulation, and the result was summed into this image. This image serves as a measure of the average integrated flux level from the noise from the continuum subtraction process. Panel (c) shows the standard deviation of the subtraction flux results shown in panel (b). The images in Figures 8(a)–(c) share the same flux scaling from 0.0 to 1.0 × 10⁻¹⁵ erg cm⁻² s⁻¹. Panel (d) presents panel (a) divided by panel (c), which is the S/N of extended H₂ line emission in the environment of AB Aur. There is low-level H₂ emission surrounding the AB Aur system at a statistically significant level of detection. The integrated line flux from this analysis is shown in Table 4. Note that no correction was made for the effects of flux attenuation from the occulting spot used for the AB Aur observation, so this measured line flux level is a lower limit to the true integrated H₂ emission.

The λ-dimension of the NIFS IFU data allows for extraction of moderate-resolution spectra, analogous to traditional long-slit spectroscopy, but located at the position of each 01 × 004 pixel. Figure 9 presents a 1D spectrum extracted in a circular aperture that spans the full IFU field for each of the seven stars that have strong H₂ emission detections (see Figure 4). The extraction aperture radii and corresponding H₂ 2.12 μm line fluxes are presented in Table 4. For AA Tau, DoAr 21, and V773 Tau, the weak H₂ emission lines are not easily detected over the bright integrated fluxes of the full-field spectra. In all other stars, H₂ emission is clearly discernible from the continuum. All spectra show 2.16 μm Brγ emission from stellar mass accretion and absorption features from species including Al, Mg, Fe, Na, Ca, and CO, which are typical photospheric lines observed in young stars with K–M spectral types. AB Aur is not presented in Figure 9. The use of the NIFS occulting spot was necessary to observe this very bright star, so data analysis required a different technique as described in Section 3.4. The line flux for AB Aur presented in Table 4 was derived by spatially integrating the continuum-subtracted H₂ line emission from Figure 8(a). This is a lower limit to the true line flux for the AB Aur system because no correction for the occulting spot was applied.

Zoom In Zoom Out Reset image size

In Figure 9, two spectra are plotted for each of the seven sources where H₂ was detected in Figure 4. (AB Aur is discussed in more detail in Section 3.4). The top spectrum of each pair was produced by integrating over the entire NIFS field of view and is labeled "full field." The bottom spectrum of each pair was produced by extracting the spectrum from a 3-pixel (012) radius aperture centered on the peak of the H₂ emission flux found in the continuum-subtracted images (see Figure 4) and is labeled "peak H₂." For GM Aur and LkHα 264, the peak flux was centered at the stellar continuum peak, so the smaller-aperture "peak H₂" spectrum presented here actually decreases the sensitivity to the H₂ line emission compared to the "full-field" spectrum. For all other systems, the 2.12 μm H₂ emission is extended, originating far enough from the bright stellar continuum such that the line emission is significantly stronger relative to the continuum flux than the emission observed in the "full-field" spectra.

The peak H₂ flux typically has between 5% and 10% of the integrated H₂ emission seen in the whole field (Table 5). In addition to the v = 1–0 S(1) feature, several other transitions of H₂ are detected in these spectra. Table 5 presents a summary of the extraction aperture size, the v = 1–0 S(1) flux level in the small extraction aperture, and the emission-line ratios of the other transitions compared to the measured v = 1–0 S(1) line. The 3σ detection limits are included as appropriate for sources that lacked detection of certain transitions. The uncertainties in the fluxes for the Q-branch features from 2.40 to 2.42 μm are large because the long on-source exposure times for the observed sources resulted in imperfect correction of narrow telluric absorption features in this spectral range. The uncertainties in the Q-branch emission fluxes were estimated by investigating the residual telluric features in the corrected spectra. The uncertainties are assumed to be lower limits. The H₂ flux for AB Aur is also included in Table 5, and the extraction aperture was centered on the discrete knot of emission to the northwest of the star. No additional H₂ features were identified in the spectrum of AB Aur. GM Aur and V773 Tau had significant telluric correction noise in the spectral region longward of 2.40 μm, which made uncertainties too large to permit reliable detections of Q-branch line emission.

In past studies, the measured v = 2–1 S(1)/v = 1–0 S(1) line ratio has been proposed or adopted to constrain H₂ emission excitation mechanisms (Black & van Dishoeck 1987; Burton et al. 1989a; Sternberg & Dalgarno 1989; Gredel 1994; Gredel & Dalgarno 1995; Maloney et al. 1996; Tiné et al. 1997; Takami et al. 2007; Beck et al. 2008; Greene et al. 2010). This diagnostic can work well to distinguish between nonthermal H₂ excitation processes such as UV pumping and fluorescence (e.g., Black & van Dishoeck 1987; Herczeg et al. 2002), which result in large electron populations of the higher vibration levels, from collisional excitation mechanisms such as shock heating in winds or jets. Typically the nonthermally excited H₂ emission regions have v = 2–1 S(1)/v = 1–0 S(1) line ratios of ∼0.25 or higher (UV excitation; Black & van Dishoeck 1987). In environments where densities are greater than or equal to a few times 10⁴ molecules cm⁻³, the gas quickly thermalizes and results in line ratios consistent with the excitation temperature in the environment. These regions may have H₂ emission stimulated by high-energy flux from the star or by shock excitation. Beck et al. (2008) used this line ratio diagnostic to create spatially resolved gas excitation maps in the environment of the T Tau triple system, and Gustafsson et al. (2010) furthered the analysis on T Tau.

The H₂ line ratios measured across the sample of stars in this study have a narrow range from 0.11 seen in AA Tau to 0.06 measured in GG Tau A. These values are more consistent with denser thermalized gas than a pure low-density excitation environment where the population of the higher vibration levels is enhanced. Assuming LTE, our measured line ratios correspond to excitation temperatures in the range of 1700–2100 K for the four systems that exhibit appreciable v = 2–1 S(1) emission. Table 6 presents line ratios and the associated excitation temperatures for LTE gas in these four systems, plus a lower detection limit on DoAr 21.

The H₂ v = 1–0 S(1) and v = 1–0 Q(3) transitions arise from the same upper state in the H₂ molecule, and the intrinsic Q(3)/S(1) line ratio of 0.7 is unaffected by physical conditions in the gas emission environment. The ratio of the flux from these transitions can be altered from the intrinsic value by line flux scattering off of dust grains, which decreases the ratio, or by attenuation from obscuring material along the line of sight to the emitting region, which increases the intrinsic ratio. Extended dust structures can cause variations in obscuration in the environments of young stars, and past analysis of IFU data suggests that significant spatial variations in dust may be revealed using the H₂ emission-line ratio (Beck et al. 2008; Gustafsson et al. 2010). Figure 10 shows the v = 1–0 S(1) and v = 1–0 Q(3) emission-line images and their ratio for AA Tau and UY Aur. The ratio images were created using only the spatial locations that had v = 1–0 S(1) line flux that was brighter than 7.5% of the peak H₂. The relative sensitivity to line flux at the central stellar positions is lower than at extended distances because of bright continuum emission, so ratio values at spatial locations where the stellar flux was brighter than 10% of the continuum peak are also not presented.

Zoom In Zoom Out Reset image size

The brightest extended knots of H₂ emission in both AA Tau and UY Aur have v = 1–0 Q(3)/v = 1–0 S(1) line ratios that are consistent within the uncertainties to the intrinsic value of 0.7 (purple or blue in the maps), implying no effects from scattering or line-of-sight extinction. The H₂ emission in AA Tau that extends to the southeast of the star has a higher line ratio, increasing to levels of ∼0.77. This ratio corresponds to a visual extinction level of ∼5 mag. This could be indicative of higher extinction along the line of sight to this H₂-emitting region, perhaps caused by obscuration from extended regions of an outer disk. The v = 1–0 Q(3) flux is relatively stronger in the spatial regions around UY Aur B compared to A, which may imply greater line-of-sight obscuration toward this known IRC. The extension to the north of UY Aur B traces an increased line ratio at a level of 0.9–1.0, which translates to strong levels of extinction of 12–18 mag in this H₂ emission region. However, the absolute v = 1–0 Q(3)/v = 1–0 S(1) line ratios typically have uncertainties of ±0.1–0.2, which corresponds to A_v differences of 7–12 mag of flux attenuation. Hence, the absolute magnitude of the A_v variation is less important than the overall relative spatial structure in these line ratio maps. DoAr 21, GG Tau, GM Aur, LkHα 264, and V773 Tau produced detectable levels of v = 1–0 Q(3) emission in the "full-field" and/or "peak H₂" spectra in Figure 9, but either the extended emission was too faint or the fluxes were too strongly affected by telluric correction uncertainties to provide a reliable line ratio image at each spatial position.

The IFU data presented in this study possess inherent uncertainties in absolute velocity calibration due to co-adding exposures collected over several nights, as described in Section 2. With the exception of UY Aur, the data have a velocity resolution accuracy of no better than 30 km s⁻¹. For UY Aur, where all of the IFU data were acquired on a single night (Table 2), the velocity accuracy is estimated to be 12 km s⁻¹. In all other systems, absolute or relative velocity structures that are less than the accuracy of ∼30 km s⁻¹ are unresolved. Clear detections of H₂ kinematics are seen in UY Aur and V773 Tau, which both exhibit velocity structure at greater than 30 km s⁻¹. Figures 11 and 12 show the kinematics of the H₂ emission seen in these two systems. In both cases panel (a) shows the barycenter velocity (flux-profile-weighted average velocity) of the H₂ emission for each spatial location that had a line flux of greater than 8σ over the noise. A more detailed description of the velocity barycenter analysis methodology for IFU data is presented in Beck et al. (2008).

Zoom In Zoom Out Reset image size

In Figure 11(b), the blueshifted emission is seen to exhibit a strong knot of H₂ to the north of V773 Tau. In Figure 11(c), the low-velocity channel map of the H₂ emission shows the gas distributed in knot-like structures all around V773 Tau, with more diffuse emission coming from gas extending to distances of ∼200 au or greater. The position of the peak in H₂ flux is coincident with the location of the companion, 02 from the primary source. In Figure 11(d), a bright knot of redshifted H₂ emission is seen to the south of the V773 Tau system. High-velocity blue- and redshifted knots of emission around V773 Tau are obvious in Figure 11, presenting evidence of a bipolar outflow from this complex young multiple system. To our knowledge, this represents the first detection of spatially extended lobes of outflow emission from the V773 Tau system.

UY Aur shows blueshifted emission to the southeast of the primary in the system (Figure 12(b)). UY Aur also shows knot-like emission and bright arcs that extend away from both stars. The brightest emission region is to the southeast of the primary, UY Aur A (Figure 12(c)), and appreciable redshifted flux is seen to the northwest of UY Aur B (Figure 12(d)). The kinematic structure in the H₂ observed around UY Aur shows a gradient of ∼20 km s⁻¹ from blueshifted H₂ to the southeast of the field of view to redshifted emission in the northwest. Interestingly, this observed velocity gradient in the UY Aur system is perpendicular to the position angle of the jets identified in Pyo et al. (2014). Moreover, multiple knots of emission from blueshifted velocities of −40 km s⁻¹ (position A, Figure 12(b)) to redshifted velocities of +10 km s⁻¹ (position C; Figure 12(d)) correspond to features observed in the [Fe ii] jets from UY Aur as measured in Pyo et al. (2014). Cross-referencing the kinematic features with the proposed system geometry from Pyo et al. (2014) clarifies the nature of the H₂ emission from UY Aur. This is discussed in detail in Section 4.1.

Zoom In Zoom Out Reset image size

In order to compare H₂ emission-line luminosity to other characteristics of the observed systems, we use information from the literature and from our K-band IFU spectra to collect and derive stellar parameters. The distances to our surveyed systems are taken from the Gaia DR2 catalog for most stars (Bailer-Jones et al. 2018). An ensemble average for the Taurus dark cloud of 140 pc is used for GG Tau A because the DR2 distance seems affected by the 025 separation multiple system. Integrated H₂ emission-line luminosities and Brγ fluxes are from this study. X-ray luminosities are from the XEST survey for all available systems (Güdel et al. 2007, 2010). X-ray measurements of AB Aur and GG Tau A came from Telleschi et al. (2007) and P. C. Schneider (2019, private communication), respectively. Visual extinctions, temperatures, and luminosities are taken from Herczeg & Hillenbrand (2014), who compiled a consistent data set for all of the systems in our survey. The luminosities for the young multiples are adjusted based on K-band flux ratios for the spatially unresolved systems from Herczeg & Hillenbrand (2014), and presented values are for the primary stars in the systems. DoAr 21 was analyzed assuming that the source is an equal-mass double star (Loinard et al. 2008). Stellar masses and ages are derived here, using these data from Herczeg & Hillenbrand (2014) and the evolutionary tracks of Baraffe et al. (2015). The more massive DoAr 21 and AB Aur systems required use of information from Feiden (2016). The results for masses and ages are very consistent with published values for each of these systems, but we have rederived the information to ensure a consistent comparison. The mass accretion rates are determined using our measured H i Brγ line flux and converting to an accretion luminosity using the methodology of Gullbring et al. (1998), and using the existing scaling relations to equate that to mass accretion (Muzerolle et al. 1998; Manara et al. 2013). All results are presented in Table 7 and used for comparison and discussion in Section 4.

In Beck et al. (2008), six stars with known Herbig–Haro outflows were surveyed for spatially extended rovibrational H₂ emission. Of those six stars, all were found to have appreciable H₂ emission that was predominantly excited by shocks in the jets or winds. In this project, we specifically targeted sources that had evidence suggestive of a more disk-like origin for the H₂ emission. In the eight systems where we measure the rovibrational H₂, strong centralized gas emission from the inner disks likely exists in all stars except for DoAr 21. AA Tau, UY Aur, and V773 Tau show shock-excited H₂ emission from known and newly revealed inner outflows.

AA Tau exhibits a slight blueshifted line H₂ line profile in its high-resolution spectrum (Table 2), and the brightest knot of emission to the southwest at 07 (100 au) distance from AA Tau lies along the axis of the known jet (Cox et al. 2013). The 02 aperture "peak H₂" spectrum presented in Table 5 and Figure 9 was extracted at this location. The measured v = 1–0 Q(3)/v = 1–0 S(1) line ratio at the knot position reveals a very low level of line-of-sight obscuration toward the extended region of the jet (Figure 10). This bright knot of jet emission also shows measurable v = 2–1 S(1) flux with a ratio of 11% compared to the 1–0 S(1) transition. For gas in LTE, this emission corresponds to an excitation temperature of ∼2200 K (Table 6). This emission at 100 au distance along the jet channel with a temperature of over 2000 K clearly points to shocks in the outflow as the responsible excitation mechanism for the H₂.

Figure 11 shows the average barycenter velocity for the H₂ emission from V773 Tau, revealing north–south bipolar knots. The knots are positioned asymmetrically around the system; the blueshifted knot is ∼06 to the north of V773 Tau A, and the southern redshifted knot is ∼11 to the south and half off of the field of view (Figure 11). It is not clear which star in this subarcsecond quintuple system might be exciting this outflow. V773 Tau shows the strongest blue- and redshifted H₂ of any of the stars in the current nine-source sample. In fact, of the 15 stars now surveyed for H₂ emission (Beck et al. 2008), V773 Tau shows the second-strongest kinematic structure in the H₂ emission profile, behind only the anomalously strong redshifted jet seen in RW Aur.

UY Aur has a known Herbig–Haro flow (HH 386; Hirth et al. 1997), with spatially resolved inner jets from both the A and B stars and a wide-angle wind from UY Aur A (Pyo et al. 2014). The collimated blueshifted jet from UY Aur B extends to the northeast toward UY Aur A, and the redshifted flow from UY Aur A extends to the southwest toward UY Aur B. The apparent outflows nearly align with the binary orientation. We clearly detect strong H₂ emission in several knots that can be traced to these known [Fe ii] jets (Pyo et al. 2014). In Figures 12(b)–(d) and Figure 14, three positions A, B, and C are designated. Position A is the location of the strongest measured blueshifted H₂ in the environment of UY Aur. Comparison with the [Fe ii] maps of Pyo et al. (2014) shows that this emission traces the inner blueshifted jet from UY Aur B. Position B traces a strong knot of slightly redshifted outflow emission, and Position C shown in Figures 12 and 14 traces strongly redshifted emission from the southwestern jet from UY Aur A. The bright arc of H₂ emission that we detect to the southeast of UY Aur A has a counterpart in the low-velocity [Fe ii] blueshifted maps presented by Pyo et al. (2014).

Zoom In Zoom Out Reset image size

The UY Aur A + B binary exhibits extensive H₂ emission that fully encompasses this young system (Figure 14(a)). Three arcs surrounding UY Aur A and B are highlighted in green, yellow, and cyan in Figure 14(a). The green arc spans a range of nearly 180°, opens toward the north−northeast, and fully encompasses the south side of UY Aur A. Pyo et al. (2014) also saw the brighter inner regions of the green arc in low-velocity [Fe ii] emission and interpreted this structure as arising from a blueshifted wind with a very wide opening angle (90° from Pyo et al. 2014) and a special viewing geometry tilted toward the observer. The yellow line overplotted in Figure 14(a) is another arc of emission around UY Aur A that is in the H₂ map. This arc starts at the northern side of UY Aur A and wraps around the east to the south−southeast of the star. This particular arc does not seem to have a direct counterpart in [Fe ii] emission measured by Pyo et al. (2014), but the overall morphology is similar to the geometry that they propose for the corresponding redshifted wide-angle outflow from UY Aur A. However, inner redshifted outflows are usually very difficult to measure because the optically thick central dust disk blocks clear detection of the emission on the far side of the disk. UY Aur A has a dense inner dust disk measured by ALMA (Tang et al. 2014). If the line traced in yellow is an arc from the redshifted wind, the fact that we see this arc initiate within ≤015 from the north of UY Aur A and wrap around to the east−southeast necessitates a special viewing geometry and a sparse amount of obscuring dust in the inner ∼20 au of the circumstellar disk. The third line overplotted in cyan in Figure 14(a) traces a similar arc of emission from UY Aur B.

The UY Aur system geometries presented in Figure 5 of Pyo et al. (2014) are updated and shown here in Figures 14(b) and (c). Figure 14(b) presents our observed view, and Figure 14(c) shows an expanded view of the geometry as viewed from the right side (consistent with Pyo et al. 2014). The location of the measured collimated jets, wide-angle wind components, and regions with strong H₂ emission are highlighted. The strongly redshifted knot C is associated with the outskirts of the jet from UY Aur A (Pyo et al. 2014). These schematic geometrical descriptions of the UY Aur system also reveal that the bright H₂ knot designated B in Figures 12 and 14 is likely associated with extended regions in the wide-angle redshifted flow from UY Aur A. We hypothesize that the extended arc-like H₂ structures seen in the environment of UY Aur A are excited in part by shocks at the outskirts of the wide-angle outflows, possibly through interaction of the multiple wide-angle flow components that are slightly inclined with respect to each other (Pyo et al. 2014). For example, the collimated blueshifted jet from UY Aur B seems like it could be interacting with the wide-angle wind from UY Aur A, producing the extended arc of emission highlighted in green in Figure 14(a). The great extent of this H₂ arc implies that there might be a wide-angle wind component from UY Aur B that is also colliding and shocking the H₂. Hence, we postulate that young star multiples have outflow–outflow interactions as an additional H₂ shock excitation mechanism that does not exist in single-star systems.

Combined with the results of Beck et al. (2008), of the 15 sources surveyed for extended H₂ emission, eight systems show strong evidence for H₂ in extended outflows. Four of these exhibit resolved kinematic structure beyond the 30 km s⁻¹ instrumental limits: RW Aur, T Tau, UY Aur, and V773 Tau. We plotted mass accretion rate, age, and stellar mass versus H₂ emission luminosity for the Beck et al. (2008) sample of six stars driving HH outflows, as in Figure 13. We find no apparent correlations. The observed anticorrelation between X-ray luminosity and H₂ emission luminosity shown in Figure 13(a) is not as strong if the six targets from Beck et al. (2008) are included in the analysis. This is perhaps because several of the targets in that survey were very young (<∼1 Myr) and/or highly obscured, which limits measurement of the X-ray luminosity.

Stars in the GG Tau A, UY Aur, and V773 Tau young star multiple systems with separations greater than 02 were spatially resolved by our study (Figure 3). DoAr 21 is a tight binary system; it was resolved into two stars in 3.6 cm radio continuum emission with a separation of 5 mas (0.6 au; Loinard et al. 2008). For the case of the sub-au separation for DoAr 21, heating from two stars in a central binary should increase the detected H₂ disk emission luminosity and extent compared to a single central star of equivalent spectral type. Interestingly, DoAr 21 exhibits no compact and central H₂ emission. It is the only system in our sample with appreciable H₂ but no evidence for an inner flux distribution suggesting emission from a central disk component. In fact, there appears to be little to no H₂ at any location within a ∼60–70 au distance from DoAr 21. All of the H₂ emission arises from the extended distribution spanning a 120° arc from the north to southwest of the central DoAr 21 binary system (Figure 4).

GG Tau A and UY Aur show H₂ morphologies with arcs of emission that extend away from inner stars in the systems (Figures 14(a) and 15(a)). As discussed in the previous section, the arcs encompassing UY Aur are seen in between stars and are likely shock excited by outflows and possibly interacting winds. We detect two arcs of emission (yellow and cyan in Figure 14(a)) that are interpreted to arise from the extended redshifted outflow from both UY Aur A and B based on the [Fe ii] kinematics and geometry found by Pyo et al. (2014). The yellow arc is seen to arise from within <015 of UY Aur A, at a location where significantly redshifted outflow emission should be obscured by the optically thick inner dust disk. Tang et al. (2014) find highly complex dynamics of the CO gas surrounding the UY Aur binary within the circumbinary ring cavity from interaction of material in Keplerian rotation with gas tracing both the mass accretion infall and outflow. Their study does not clearly resolve the gas in the inner ∼1'' of UY Aur A. While we find that shock excitation in the wide-angle redshifted outflows is the most likely explanation for these two arcs of emission, alternative excitation of gas in mass accretion infall shocks cannot be firmly excluded based only on the morphology of the H₂. Streamers of infalling material would also appear to be slightly redshifted at a ∼20–30 km s⁻¹ level (Pyo et al. 2014) and might be observable at <015 distances from UY Aur A in the foreground of the circumstellar dust disk.

In the case of the complex subarcsecond triple system GG Tau A (Di Folco et al. 2014), the H₂ emission arcs encompass the stars in this young system (Figure 15(a)). The brightest region of H₂ emission in the environment of GG Tau A is ∼40 au to the northeast of the stars. This H₂ emission has an estimated LTE excitation temperature of ∼1700 K, estimated from the detected level of v = 2–1 S(1) line emission. Based on models for H₂ excitation in circumstellar disks (Nomura et al. 2007), at a distance of ∼40 au from the stars, this temperature is on the high side to be excited only by flux from the central stars. However, this region of strongest H₂ gas emission is not in a disk; it is located in a dynamically unstable region that must be continuously replenished, or it will be cleared on ∼100 yr timescales. Streamers of infalling material that transfer mass from a circumsystem disk to the inner stars are predicted by hydrodynamical theories of binary star evolution (Artymowicz & Lubow 1996). Streams of CO gas that connect the outer GG Tau A circumsystem ring to the inner H₂ gas distribution have been measured (Dutrey et al. 2014). GG Tau A is a complex triple system, and Beck et al. (2012) postulated that this extended distribution of H₂ emission might also be excited in part by low-velocity shocks from accretion infall as material streams inward from the massive circumsystem ring that surrounds the stars. The accretion infall shock velocity would result from a combination of the gas freefall velocity onto the system (∼15 km s⁻¹) with a component from the Keplerian motion of material in the inner system. The presence of the sub-40 au triple system would further complicate gas kinematics of the regions surrounding the stars. Shock velocities in the lower 20–30 km s⁻¹ range of models (Le Bourlot et al. 2002) could explain the H₂ in the GG Tau A system (Beck et al. 2012), but these velocities are quite large compared to infall+Keplerian kinematic measurements of cooler, denser circumbinary disk gas (e.g., Dutrey et al. 2014, 2016; Tang et al. 2014). Figure 15(a) highlights in green the location of two of the obvious extended arcs of emission that encompass the GG Tau A triple star; the eastern arc traces the inner edge of the CO streamer (Dutrey et al. 2016).

Zoom In Zoom Out Reset image size

V773 Tau is a putative young subarcsecond quintuple (Ghez et al. 1993; Leinert et al. 1993; Duchêne et al. 2003; Boden et al. 2012). V773 Tau A was found to be a double-lined optical spectroscopic binary (V773 Tau Aa+Ab SB2; Welty 1995). Orbital monitoring and dynamical mass fits of nearby V773 Tau B suggest that it is also an ∼au-scale binary (V773 Tau Ba+Bb; Boden et al. 2012). In our continuum image of V773 Tau in Figure 3, The V773 Tau A + B quadruple system is the spatially unresolved brighter point-like component. Boden et al. (2012) showed that the A + B separation at the time of our 2009 observation was ∼50 mas, and so it was below our spatial resolution limit. The 025 separation companion that we do spatially resolve in Figure 3 is V773 Tau C (designated "D" in the discovery paper by Duchêne et al. 2003 but later renamed by Boden et al. 2012). V773 Tau C is an IRC; it is the brightest star in the system at wavelengths longward of 4 μm (Duchêne et al. 2003). Figure 15(b) shows the image of the low-velocity (v = −45–15 km s⁻¹) H₂ emission from V773 Tau. The brightest region of H₂ emission is seen at the position of V773 Tau C. The blue- and redshifted knots in the bipolar outflow are seen to the north and south of the stars. The two green lines in Figure 15(b) trace distributions of brighter H₂ emission arcs that are near the stellar rest velocity. These arcs seem to partially encompass the stars in the V773 Tau system in a direction that is perpendicular to the outflow knots. The morphology of the two arcs resembles a circumsystem distribution of material that encircles the stars from the east to the west. However, it is unfortunately not possible to discern from the image whether these two distributions of H₂ emission are related or independent, or what the excitation mechanism might be.

When combined with the six sources from Beck et al. (2008), 15 young stars have been surveyed with Gemini+NIFS for spatially resolved rovibrational molecular hydrogen. Of these 15, 7 are young single-star systems (AA Tau, AB Aur, DG Tau, HL Tau GM Aur, LkCa 15, and LkHα 264) and 8 are young multiple stars (DoAr 21, GG Tau A, HV Tau C, RW Aur, T Tau, UY Aur, XZ Tau, and V773 Tau). Figure 16 investigates the H₂ extent and brightness across the full sample of stars. This shows the H₂ line luminosity versus this H₂ emission area measure for single stars (triangles), resolved multiple stars (diamonds), and sub-au multiples (asterisks), and the two high-A_v sources are also encompassed by a square. The H₂ sky emission area used in Figure 16 shows the region of the spatial field (in au²) that exhibits greater than 5% of the H₂ flux compared to peak line emission pixel position. This area measure is calibrated for distance in the stellar frame (Table 7) and is not affected by the S/N of the observation because all measurements were sensitive to less than 5% of the peak flux. Overplotted as dashed–dotted lines are the averages and ranges for single stars and multiples; the lines cross at the average position, and the range is defined by the span of the triangle and diamond points, respectively. AB Aur, HL Tau, and HV Tau C were omitted for the average and ranges because of the occulting spot used for the observations (AB Aur, HL Tau), and the high source extinction (HL Tau, HV Tau C) preferentially increases sensitivity to extended H₂ emission. Figure 16 shows that the average H₂ emission area is ∼3 times greater in the young multiples versus single stars, and the average v = 1–0 line luminosity is nearly a factor of 10 higher. DoAr 21 has an H₂ line luminosity and spatial extent that are similar to the measured singles rather than the multiple systems.

Zoom In Zoom Out Reset image size

The young star multiple systems surveyed thus far have stronger H₂ emission line luminosity over a greater overall spatial area than the single-star systems. This is naturally expected for spatially resolved binaries or higher-order multiples because the two (or more) disk+outflows have greater combined flux and span a larger overall area on the sky compared to single-star systems (e.g., Figure 1). Interpretation of results for GG Tau A and UY Aur suggests that at least two additional H₂ gas excitation mechanisms may exist solely in young multiple systems: shock excitation from system dynamics and mass accretion infall (GG Tau A), and shocked arcs from interacting jets and wide-angle outflows in slightly misaligned systems (UY Aur).

Of the nine stars surveyed for this project, we find that LkCa 15 and DoAr 21 exhibit no central H₂ emission associated with the stellar point source constraining molecular gas in the inner disks. All other systems have appreciable central H₂ flux that likely has some contribution from disk gas excited by central emission from the star(s). AA Tau, GG Tau A, UY Aur, and V773 Tau also have shock excitation mechanisms stimulating spatially extended H₂, which makes line ratio and spatial morphology analysis of the central disk emission difficult. However, AB Aur, GM Aur, and LkHα 264 exhibit near-IR H₂ that is most consistent with a pure disk origin. Emission in these systems is characterized by spatially compact and centrally dominated line flux with no measurable kinematic structure and no evidence for extended discrete knots or high LTE gas temperatures suggesting shock excitation or an outflow origin.

Figure 17 shows the log-scaled continuum-subtracted H₂ images of GM Aur, AB Aur, and LkHα 264. The cyan ellipse in the GM Aur view (panel (a)) presents the orientation of circumstellar disk material sampled by the 0.9 mm dust continuum emission, shown here in the location of the 50× rms dust contour from Figure 1(a) of Macías et al. (2018). The near-IR H₂ emission from GM Aur is spatially compact, and the majority of the flux arises from within 02 (∼32 au) of the central star. GM Aur has an inner cavity in the dust disk within 35 au (Hughes et al. 2009; Macías et al. 2018), and the H₂ gas fills this region. The low-level H₂ from GM Aur shows nonaxially symmetric structures with a slight overall north–south extension.

Zoom In Zoom Out Reset image size

Two green lines overplotted on the log-scaled H₂ image of AB Aur (Figure 17(b)) highlight the position of the CO gas spirals measured by Tang et al. (2017). These spirals exist within a dust-cleared ring that encircles AB Aur at a distance of ∼120 au. The ¹²CO 2–1 emission spirals trace optically thick cool disk material at a temperature of ∼140 K (Tang et al. 2017). We see spatially coincident H₂ emission that traces warmer gas in the lower-density upper layers of the spirals. Hence, the spiral-shaped density enhancements seen around AB Aur exist in multiple gas species that trace a range of scale heights in the disk. Moreover, the low-level H₂ flux seen around AB Aur also shows arc-like spiral extensions (e.g., to the north in Figure 17(b)), tracing sculpted gas at greater distances than revealed in the CO. Tang et al. (2017) postulate that the spiral arms in the dust-cleared AB Aur disk cavity might be shaped by tidal disturbances in the gas imparted by two planets at separations of ∼30 au and ∼60–80 au.

The disk emission from LkHα 264 is spatially compact, and like GM Aur, the majority of the flux arises from within 02 (∼32 au) of the central star. The overall emission is slightly asymmetric toward the west, and two linear extensions of H₂ emission are seen to ∼06 (150 au) west of the star. The origin of these extensions is not clear, the H₂ shows no obvious kinematic characteristics of shock-excited gas (Itoh et al. 2003; Carmona et al. 2008b), and LkHα 264 is not known to have an extended inner jet.

Nomura & Millar (2005) constructed a self-consistent density and temperature model that accurately reproduces molecular hydrogen emission in a young star disk that has strong UV excess radiation. The model was constructed for a generic T Tauri star of mass 0.5 M_⊙, radius 2.0 R_⊙, and temperature 4000 K. Their Figure 10 presents the v = 1–0 S(1) flux model calculated as a function of distance from 0.1 to 100 au from the central star. Figure 18 plots the average measured flux as a function of distance from GM Aur and LkHα 264, with the Nomura & Millar (2005) model scaled to the emission level at 20 au and overplotted. Nomura & Millar (2005) find that the physical and chemical structure of protoplanetary disks is affected by the dust and that H₂ emission in the central region peaks at a ∼20 au distance from the star. We do not see an inner decline in H₂ emission within 20 au in the disk of GM Aur, and LkHα 264 is too distant to resolve the inner 20 au region. Beyond 20 au, the average radial H₂ flux structure in GM Aur and LkHα 264 closely resembles the Nomura & Millar (2005) model. It should be noted that the inner H₂ emission from GM Aur arises from within the dust-cleared inner disk, and the Nomura & Millar (2005) model may not be directly applicable inside the cavity. Hoadley et al. (2015) found that UV H₂ emission in transition disk systems, such as GM Aur, extended to four times greater radii than nontransition disks, and that all of the warm H₂ came from within the inner dust-cleared cavity. We see rovibrational H₂ that extends beyond the ∼40 au radius inner dust-cleared ring of GM Aur Macías et al. (2018), confirming that the near-IR H₂ traces gas to larger radii than the UV transitions.

Zoom In Zoom Out Reset image size

Expanding on the work of Nomura & Millar (2005), Nomura et al. (2007) updated the H₂ disk models to include effects of disk evolution and dust grain growth. As the dust particles evolve in protoplanetary disks, the gas surface temperature drops because grain photoelectric heating becomes inefficient. The effect of this grain growth manifests itself in an increase in nonthermal pumping versus the collisionally dominated gas in LTE, and this can be measured as an increase in the v = 2–1 S(1)/v = 1–0 S(1) near-IR H₂ line ratio. Particularly, models with significant grain growth (to 10 cm) are traced by near-IR H₂ gas line ratios of up to 0.15 (Nomura et al. 2007). We do not detect the v = 2–1 S(1) emission in GM Aur, AB Aur, or LkHα 264, so Figure 19 plots the v = 2–1 S(1)/v = 1–0 S(1) line ratio limit for these three systems from the integrated disk flux apertures (shown on the x-axis). Overplotted are the line ratio calculations for Nomura et al. (2007) "Model A" models that incorporate X-ray plus UV flux for small (10 μm), medium (1 mm), and large (10 cm) dust grains. The observed line ratio limits rule out significant grain growth to 10 cm grains for the LkHα 264 system. The detection limits on the v = 2–1 S(1) line flux for GM Aur and AB Aur are not restrictive enough to measure any grain growth in the inner dust-cleared regions of these systems, where the effect might be most prominent. Additionally, an increased v = 2–1 S(1)/v = 1–0 S(1) line ratio in the inner regions could also be misinterpreted as shock-excited emission, such as from an inner disk wind, unless the ratio is greater than ∼0.25 and more indicative of a nonthermal origin.

Zoom In Zoom Out Reset image size

The v = 2–1 S(1)/v = 1–0 S(1) line ratio of 0.046 measured in the 30–195 au aperture for the disk of LkHα 264 places a temperature upper bound of 1650 K for the LTE gas. Prior studies have estimated molecular hydrogen disk masses under the assumption that the bulk of the rovibrational H₂ emission arises from a collisionally dominated region with an average gas temperature of ∼1500 K (Bary et al. 2003, 2008; Itoh et al. 2003; Carmona et al. 2008b). Our results and line ratio limits are consistent with these past assumptions. However, proper calibration of the spatially resolved molecular hydrogen disk mass structure in GM Aur, AB Aur, and LkHα 264 would benefit from treatment with an updated 3D model of these disks. Particularly, the H₂ emission from the inner disk regions of GM Aur and AB Aur arises from dust-cleared cavities, and existing models use dust as a dominant disk opacity source that affects the inner radiation field and characteristics of the emitting gas. Further modeling of the H₂ in these disks is a topic for a follow-up project.

4.4.1. AA Tau

4.4.2. AB Aur

4.4.3. DoAr 21

4.4.4. GG Tau A

4.4.5. GM Aur

4.4.6. LkCa 15: A Null H₂ Detection

4.4.7. LkHα 264

4.4.8. UY Aur

4.4.9. V773 Tau

The IFU data presented in this study are complex. Figure 4 presents a single analysis method across all T Tauri star systems, to investigate extended H₂ emission per spatial element in the data cube. The 3D IFU data structure allows for tailored, specific analysis methods to investigate the sensitivity to measured H₂ depending on characteristics of the system (e.g., Figure 8 for AB Aur). There are a multitude of ways to analyze the IFU H₂ line emission data, and each young star system presented and discussed here is worthy of it is own full-length article that delves into significantly more detail, analysis, and modeling. As a result, we recognize that these IFU data need to be publicly available for future complementary studies and comparison. The raw NIFS data presented in this study are available in the Gemini Observatory data archive. However, the integral field spectroscopic observing modes require training and careful attention to process the data correctly, and it is very time-consuming to do so. As a result, the v = 1–0 S(1) H₂ emission-line maps, 2.1 μm continuum images, and the fully reduced and flux-calibrated 3D IFS data cubes (∼2.01–2.45 μm) used for this project are all available for public download at http://www.stsci.edu/~tbeck/data/2019/NIFSH2. These files and any of the other line emission maps presented in this study can also be made available upon request to the first author. The emission-line maps that we present here provide a valuable demonstration of the types of data sets expected on CTTSs from the James Webb Space Telescope Near-IR Spectrograph (NIRSpec) and Mid-IR Instrument (MIRI) IFU observing modes.