DISCOVERY OF THE Y1 DWARF WISE J064723.23−623235.5

Why is the discovery of a single new Y dwarf important? First, promptly announcing new Y dwarf discoveries means that the astronomical community can perform various follow-up investigations while the Spitzer Space Telescope and Hubble Space Telescope (HST) are still active. Ground-based observatories are hindered by the extreme faintness of these objects below 2.5 μm (see Figure 12 of Kirkpatrick et al. 2012) and by the opaqueness of Earth's atmosphere at thermal and mid-infrared wavelengths where these objects emit most of their light. Space-based facilities do not suffer from these limitations and can be used, for example, to obtain accurate distances via precision astrometry (Beichman et al. 2013; Marsh et al. 2013; Dupuy et al. 2013), to search for transiting exoplanets that have brown dwarfs as their host "suns" (Triaud et al. 2013), and to probe the evolution of features in the cloud deck via photometric monitoring (Buenzli et al. 2012; Apai et al. 2013; Heinze et al. 2013). Unfortunately, these orbiting facilities are nearing the end of their lifetimes, so quick publication is critical.

Second, only sixteen Y dwarfs are known (Cushing et al. 2011; Kirkpatrick et al. 2012; Tinney et al. 2012; Liu et al. 2012; Cushing et al. 2013), along with another assumed to be a Y dwarf but still lacking spectroscopic confirmation (Luhman et al. 2011). When in the realm of small number statistics, the discovery of a single, new object in the solar neighborhood is still significant, as this has a substantial impact on our understanding of the Y dwarf space density. An accurate space density is crucial because it is these coldest objects that can reveal the low-mass cutoff for star formation (see, for example, the simulations in Burgasser 2004). Gaining an accurate count of the nearby Y dwarf population will enable us to study how abruptly, or not, the observed space density below 300 K drops (Figure 14 of Kirkpatrick et al. 2012), which in turn will shed light on whether we have found the end of star formation or whether far colder Y dwarfs, beyond the reach even of NASA's Wide-field Infrared Survey Explorer (WISE; Wright et al. 2010), are expected. Stated another way, does the low apparent space density of the latest Y dwarfs argue that we have already found star formation's lowest-mass by-products; do these objects represent another population entirely, such as rogue planets (Beichman et al. 2013; Sumi et al. 2011); or does the diversity of the spectral and photometric properties of the Y dwarfs merely reflect the empirical dispersion of cold, chemically complex atmospheres?

To this end, we present the discovery of a new Y dwarf whose follow-up data indicate that it is one of the coldest and lowest-mass solivagant objects yet found outside our solar system.

Located in the constellation of Pictor, WISE J064723.23−623235.5 (hereafter abbreviated as WISE 0647−6232) was first imaged by WISE on 2010 May 9 and uncovered as a very cold brown dwarf candidate on 2010 June 17 after preliminary data processing. The goal of this candidate selection, described in detail in Kirkpatrick et al. (2011), was to identify brown dwarf candidates of type T6 or later by searching the database for sources with a W1–W2 color greater than 2.0 mag (see Figure 1 of Kirkpatrick et al. 2011 or Figure 1 of Mace et al. 2013a). WISE 0647−6232 easily satisfies this criterion with its value of W1−W2 > 3.77 mag (WISE All-Sky Source Catalog, Cutri et al. 2012; see Table 1). The lack of a detection at other WISE bands as well as at the shorter wavelengths imaged by the Two Micron All-Sky Survey (2MASS; Skrutskie et al. 2006) and the Digitized Sky Survey, further cemented this object as an excellent cold brown dwarf candidate. A finder chart is shown in Figure 1.

Zoom In Zoom Out Reset image size

3.1. Spitzer/IRAC

3.2. Magellan/PANIC

3.3. Magellan/FourStar

3.4. Magellan/FIRE

On 2013 March 24 (UT), during a night of exceptional seeing (<05) at Las Campanas, we attempted a spectroscopic observation of WISE 0647−6232 using the 6.5 m Magellan Baade Telescope with the FIRE spectrograph (Simcoe et al. 2013) in long-slit prism mode, which gives a resolving power of R = 400. Four 600 s integrations were taken, and these were nodded along the 06 slit in an ABBA pattern. The airmass range was 1.51–1.67. Data processing and analysis procedures followed those outlined for the Y dwarf WISE J163940.83−684738.6 in Tinney et al. (2012), with the only difference being that a standard extraction was used rather than that paper's detailed removal of the contamination of a nearby star. Flux and telluric calibration was performed using an observation of the A0 V star HD 51139 obtained at a similar airmass. Figure 2 shows the resulting spectrum for WISE 0647−6232. Although the quality of these data is too low to permit accurate classification, they are still useful as confirmation that the source is a late-type brown dwarf because the characteristically narrow opacity holes at Y, J, and H bands are the only wavelength regions where flux is detected. These data also demonstrate what is possible (at J ≈ 23 mag) in ideal atmospheric conditions using a large-aperture ground-based telescope with a suitable low-resolution infrared spectrograph and 40 m of on-source integration.

Zoom In Zoom Out Reset image size

3.5. HST/WFC3

Higher signal-to-noise spectroscopy was acquired using the Wide Field Camera 3 (WFC3) on board HST as part of Cycle 20 program 12970 (PI: Cushing). WFC3 employs a 1024 × 1024 HgCdTe detector with a plate scale of 013 pixel⁻¹ to image a field of view of 123'' × 126''. We first obtained pre-images of each field so that we could determine specific roll angles for the slitless grism images. This was done to ensure that the object's spectrum would not be contaminated by those of field sources. Our pre-image for WISE 0647−6232 was acquired on 2013 February 11 (UT) in the F125W filter, which is a close approximation to the J band. Integrations at four different dither positions were taken in a single orbit with the SPARS50 sampling sequence and thirteen MULTIACCUM samples for a total exposure time of 2412 s. Using this image, we computed roll angles for the grism observations.

These grism observations, acquired on 2013 May 13/14 (UT), consisted of direct images in F125W (SPARS25 sampling with eight to twelve MULTIACCUM samples) at three dithered positions in addition to dispersed images in G141 (SPARS100 sampling with thirteen MULTIACCUM samples) at four dithered positions. The total integration time was 734 s for the direct images and 9624 s for the dispersed images. The G141 grism was used for the spectroscopic observations because it covers the 1.1–1.7 μm range (resolving power of R ≈ 130) that samples the diagnostic J- and H-band peaks in the spectrum of cold brown dwarfs.

Photometry was measured on the drizzled images, and magnitudes were measured on the Vega system. Photometric reductions followed the recipe described in Cushing et al. (2011). Our measured magnitude from these images⁸ is given in Table 1. Spectroscopic reduction generally followed the recipe outlined in Cushing et al. (2011). However, a second background subtraction was required, so a two-dimensional surface was fit after masking pixels containing flux from the target, and this was subtracted from the drizzled postage stamp around the target. The spectrum was then extracted using custom IDL code using a 1'' aperture. Flux calibration was accomplished using the sensitivity curves provided by the Space Telescope Science Institute (STScI). The final flux-calibrated spectrum is shown in Figure 3. This spectrum has been used to synthesize a J−H color using the MKO J- and H-filter profiles as described in Cushing et al. (2005), resulting in (J − H)_MKO = −0.75 ± 0.10 mag (Table 1). The HST spectrum cuts off shortward of the long-wavelength edge of the H_MKO bandpass, so we have assumed that the flux is zero here, as is the case for Y dwarfs and late-T dwarfs with spectra covering the full range. The uncertainty was derived by measuring the color for a suite of a Monte Carlo simulated spectra derived from the measured errors in each spectral bin.

Zoom In Zoom Out Reset image size

The J3 − J2 color of −1.21 ± 0.20 mag from Table 1 falls blueward of the J3 − J2 = −0.5 mag limit that Tinney et al. (2012) cite as strong evidence for significant methane absorption. The Magellan/FIRE spectrum confirms this, and strongly hints that the object is a late-T or early-Y dwarf. The J3 −W2 color of 7.19 ± 0.12 mag (Table 1) implies a type of ∼Y1 using the J3 −W2 versus spectral type relation shown in Figure 4 of Tinney et al. (2012).

The actual classification of WISE 0647−6232 is determined via by-eye comparison of the HST/WFC3 spectrum to the Y0 standard WISEP J173835.52+273258.9 of Cushing et al. (2011) and the tentative Y1 standard WISE J035000.32−565830.2 of Kirkpatrick et al. (2012), as shown in Figure 3. As further illustration, the spectrum is also compared to the Y0.5 dwarf WISE J154151.65−225024.9 and the ⩾Y2 dwarf WISE J182831.08+265037.7, whose classifications also come from Kirkpatrick et al. (2012). As the figure shows, the J-band peak more closely matches the Y1 standard than it does the Y0 standard, and overall is also narrower than the J-band peak of the Y0.5 dwarf. Given the close match to the overall J-band spectrum of the Y1 standard, we classify WISE 0647−6232 as a Y1 dwarf, where this classification carries the usual uncertainty of a half subclass. We have also computed the J-narrow index of Mace et al. (2013a) using the methodology described in that paper and find J-narrow = 0.69 ± 0.07. This value is lower than all of the normal Y0 dwarfs in Table 7 of Mace et al. (2013a) and is most like the J-narrow = 0.68 ± 0.10 value of WISE J035000.32−565830.2, also classified as Y1. Thus, this index quantifies what can be seen qualitatively by eye.

We measured preliminary trigonometric parallax and proper motion values using our available imaging. The resultant astrometry is listed in Table 2 and discussed further below.

The HST images consist of the F125W images discussed in Section 3.5 as well as F105W images that will be discussed in a future paper (A. Schneider et al., in preparation). The STScI's "AstroDrizzle" pipeline was used to process these data to produce final mosaicked images. The pipeline corrects for the geometric distortion of the WFC3-IR camera to a level estimated to be <8 mas (Kozhurina-Platis et al. 2009), which is of the same order or less than the extraction uncertainties of the faint target. Sources were extracted using the Gaussian-fitting IDL FIND routine to determine centroid positions. The FWHM for these undersampled data is ∼2 pixels (026), which is consistent with STScI analyses (Kozhurina-Platis et al. 2009). The images were placed onto a common (α, δ) origin using averaged positions for over 300 stellar-like objects within 45'' of the Y dwarf. For nine objects seen in all epochs and brighter than F125W = 20 mag, the single axis positional dispersion is 4 mas. For nine unconfused, stellar objects in the range 23.0 < F125W < 23.8 mag straddling the F125W = 23.4 mag brightness of WISE 0647−6232, the single-epoch positional repeatability is 15 mas. These measures of the positional dispersion appear, however, to be conservative since the repeatability of the one F125W and two F105W observations taken over the two day period in 2013 May (MJD ≈ 56426) is only 4 mas. We therefore take a compromise value of 10 mas for the single-epoch HST uncertainties.

Seven separate Spitzer observations—one at ch1 and six at ch2—were taken as described in Section 3.1. We analyzed pBCD mosaics from the Spitzer Science Center (SSC) to make astrometric measurements. We combined all epochs of ch1 and ch2 to a common reference by averaging together the positions of 85 bright sources within ∼60'' of WISE 0647−6232 and calculating small offsets from one epoch to the next to register all frames to the average value. The largest offsets were of order 200 mas but were typically much smaller, around 50 mas. The common origin for these 85 sources was determined with a precision of (Δα = −1.2 ± 5, Δδ = 0.2 ± 0.2) mas. For a single sighting of a source brighter than ch2 < 17 mag, the positional repeatability is 60 mas and increases as expected at fainter magnitudes. These values are less than the quoted 100 mas optical distortions cited by the SSC,⁹ in part because we are confining our observations to a small central part of the IRAC arrays.

As a final step in preparing the positional data for astrometric analysis, we shifted the positions onto a common reference frame by comparing twelve objects seen in common in the 60'' field around the brown dwarf, taking care to exclude sources that were obviously extended or double as seen in the higher-resolution HST images. After correction of (α, δ) offsets <100 mas, the frame pairs share a common origin with an accuracy of (σ_α, σ_δ) = (10, 15) mas, which is negligible compared to the extraction uncertainties of the Spitzer data. The Spitzer and WISE astrometric frames (Cutri et al. 2012) are tied to the 2MASS frame, which has an overall absolute accuracy of 80 mas (Skrutskie et al. 2006). The absolute positional accuracy is not, however, critical to the astrometric solution described below.

As described in Beichman et al. (2013), the astrometric observations are fitted to a model incorporating proper motion and parallax (Smart 1977; Green 1985):

$$\begin{eqnarray} \alpha ^\prime \equiv &\, \alpha _0+\mu _\alpha (t-T_0)/\cos (\delta ^\prime)\nonumber\\ \delta ^\prime \equiv &\, \delta _0+\mu _\delta (t-T_0) \end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray} \alpha (t)&=&\, \alpha ^\prime +\pi (X(t) \sin \,\alpha ^\prime - Y(t) \cos \,\alpha ^\prime)/\cos \,\delta ^\prime \nonumber\\ \delta (t)&=&\, \delta ^\prime +\pi (X(t) \cos \,\alpha ^\prime \sin \,\delta ^\prime\nonumber\\ && + Y(t) \sin \,\alpha ^\prime \sin \,\delta ^\prime -Z(t) \cos \,\delta ^\prime), \end{eqnarray} \tag{ 2 }$$

where (α₀, δ₀) are the source position for equinox and epoch T₀ = J2000.0, (μ_α, μ_δ) are the proper motion values in the two coordinates in arcsec yr⁻¹, and π is the annual parallax in arcsec. The coefficients X(t), Y(t), and Z(t) are the Cartesian coordinates, relative to the Sun, of WISE or HST in Earth orbit or Spitzer in its Earth-trailing orbit. Values of X, Y, Z for the Earth-orbiting observatories are taken from the IDL ASTRO routine XYZ. Spitzer values of X, Y, Z are obtained from the image headers provided by the SSC. Equations (1) and (2) are solved simultaneously using the Mathematica routine NonLinearModelFit incorporating appropriate uncertainties for each data point. The fit to the combined HST and Spitzer data sets yields a good solution for proper motion and parallax (Table 3) with χ² = 16.5 with 21 degrees of freedom; this can be compared to a model incorporating only proper motion, which yields χ² = 247 with 22 degrees of freedom. A model incorporating a non-zero parallax is thus strongly preferred and the slightly low reduced χ² for this case indicates that our estimates of the positional uncertainties are, as suggested above, conservative. The resultant parallax is 115 ± 12 mas, which corresponds to a distance of 8.71 ± 0.88 pc, and the resultant proper motion is 387 ± 25 mas yr⁻¹. Table 3 lists these values along with absolute magnitudes derived from the photometry in Table 1. Figure 4 shows the motion of WISE 0647−6232 on the sky.

Zoom In Zoom Out Reset image size

Table 3. Distance, Kinematics, and Absolute Magnitudes for WISE 0647−6232

Parameter	Value
α (J2000.0 at epoch 2000.0)	06h47m232270 ± 00103
δ (J2000.0 at epoch 2000.0)	−62°32'39744 ± 0295
μα ('' yr−1)	0.007 ± 0.012
μδ ('' yr−1)	0.387 ± 0.022
π ('')	0.115 ± 0.012
Distance (pc)	8.7 ± 0.9
Vtan (km s−1)	16 ± 2
Reduced χ2 (w/parallax)	0.79
Reduced χ2 (w/o parallax)	11.2
MJ, MKO (mag)	22.95 ± 0.35
MJ2 (mag)	23.99 ± 0.28
MJ3 (mag)	22.81 ± 0.24
MF125W (mag)	23.74 ± 0.24
MH, MKO (mag)	23.70 ± 0.36
Mch1 (mag)	18.19 ± 0.24
Mch2 (mag)	15.37 ± 0.23
MW2 (mag)	15.62 ± 0.24

Download table as:  ASCIITypeset image

Using the J_MKO − J3 = 0.14 ± 0.25 mag correction derived by Tinney et al. (2012) for Y dwarfs, we find that WISE 0647−6232 has J_MKO − W2 = 7.33 ± 0.28 mag and J_MKO − ch2 = 7.58 ± 0.27. The only other confirmed Y dwarf with redder colors is the enigmatic ⩾Y2 dwarf WISE J182831.08+265037.7 (Cushing et al. 2011; Kirkpatrick et al. 2012; Beichman et al. 2013), with J_MKO − W2 = 9.18 ± 0.36 mag and J_MKO − ch2 = 9.25 ± 0.35 mag (Table 2 of Kirkpatrick et al. 2012). One other object,¹⁰ WD 0806−661B (Luhman et al. 2011), is almost certainly a Y dwarf but lacks a confirming spectrum due to its faint apparent magnitude; it has J_MKO − ch2 > 7.06 mag (Luhman et al. 2012), so it may also be redder than WISE 0647−6232. The ch1 − ch2 color of 2.82 ± 0.09 mag is redder than both WISE J182831.08+265037.7 (2.60 ± 0.03 mag; Kirkpatrick et al. 2012) and WD 0806−661B (2.77 ± 0.16 mag; Luhman et al. 2011), although not as red as that of other Y dwarfs, such as the Y1 dwarf WISE J035000.32−565830 (3.25 ± 0.10 mag; Kirkpatrick et al. 2012) and the Y0 dwarf WISE J220905.73+271143.9 (3.08 ± 0.09 mag; Cushing et al. 2013). As Figure 8 of Kirkpatrick et al. (2012) shows, there is a significant scatter of ch1 − ch2 colors within the Y0 and Y1 subtypes, perhaps indicating a range of varying physical conditions (gravity, metallicity, cloud properties, etc.) within each group.

Figure 5 shows the trend of absolute W2 magnitude as a function of spectral type using previously published values from Dupuy & Liu (2012), Tinney et al. (2012), Beichman et al. (2013), and Marsh et al. (2013). The M_W2 value for WISE 0647−6232 falls ∼1.5 mag fainter than that of the latest, well measured T dwarfs on the plot and is also fainter by ∼0.4 mag than the Y0 dwarf WISE J163940.83−684738.6 (Tinney et al. 2012) and ∼1.6 mag fainter than the ⩾Y2 dwarf WISE J182831.08+265037.7 (Beichman et al. 2013). There are three Y0 and Y1 dwarfs from Marsh et al. (2013) with fainter absolute W2 magnitudes, but all of these have error bars of ∼0.5 mag or higher. Thus, WISE 0647−6232 is one of the intrinsically dimmest, solivagant, star-like bodies yet identified outside our solar system.

Zoom In Zoom Out Reset image size

What do spectroscopic models suggest about the physical parameters of WISE 0647−6232? Before answering this question quantitatively, it is illuminating to learn first what the models suggest about the qualitative trends. For this purpose, we use the BT-Settl models from the Lyon group (Allard et al. 2003), which are valid down to effective temperatures of 400 K (corresponding to early-Y dwarfs; Cushing et al. 2011) and include gravitational settling of grains from the photosphere. Figure 6 shows models in the range 400 K <T_eff < 700 K for a fixed value of log(g) = 4.5, where g is in units of cm s⁻². These models suggest that the narrowness of the J- and H-band peaks can be used as a proxy for the brown dwarf's effective temperature. However, unlike stars on the main sequence, brown dwarfs at fixed effective temperature may span a vast array of masses, so other parameters must also be considered.¹¹ Figure 6 shows the variation in spectral features over the range 3.5 < log(g) < 5.0 for a fixed value of T_eff = 400 K. Here, log(g) can be used as a proxy for mass. Note that even for a fixed value of T_eff there is variation in the widths of the J- and H-band peaks. Moreover, there are also noticeable shifts in the wavelength of these peaks,¹² with the wavelengths of the peaks shifted slightly redward for higher values of log(g). The most noticeable effect, though, is the modulation of the J − H color¹³: lower values of log(g) correspond to significantly bluer J − H colors because the H-band peak is much more strongly muted relative to the J-band peak.

Zoom In Zoom Out Reset image size

WISE 0647−6232 is the bluest Y dwarf in implied J − H color of all the Y dwarfs shown in Figure 3 and has as narrow a J-band peak as any Y dwarf yet observed. Qualitatively, then, we would expect WISE 0647−6232 to have a low effective temperature and low value of log(g) relative to other known Y dwarfs. Figure 7, showing the best by-eye fit to the BT-Settl models, confirms this. The best fit implies T_eff < 400 K (the J-band peak is narrower than any of the models in the grid) with a relatively low value of log(g) = 4.0. Evolutionary models by Burrows et al. (2003) suggest a mass of ∼5 M_Jup regardless of the exact value of T_eff.

Zoom In Zoom Out Reset image size

The BT-Settl models represent an earlier generation of theoretical spectra lacking sulfide and chloride clouds now known to be important at low effective temperatures (Morley et al. 2012). What do these new models suggest about the physical parameters of WISE 0647−6232? Like the BT-Settl grid, the grid presented in Morley et al. (2012) does not probe below an effective temperature of 400 K, but these models were extended down to 300 K in Leggett et al. (2013). Using Figures 4 and 5 of Leggett et al. (2013), our measured absolute values of M_J = 22.95 ± 0.35 and M_W2 = 15.62 ± 0.24 mag (Table 3) suggest a value of T_eff ≈ 350 K for WISE 0647−6232. The measurement of log(g) ranges from 4.0 dex (comparing to the upper panel of Figure 5 in Leggett et al. 2013) to 5.0 dex (comparing to the upper right panel of Figure 4 in Leggett et al. 2013). Comparison to the evolutionary models in Saumon & Marley (2008) suggests a mass of ∼5 M_Jup—similar to the conclusion reached using the BT-Settl spectra and Burrows et al. (2003) evolutionary models—and age of ∼100 Myr for the log(g) = 4 case. However, higher values of the mass and older ages are also indicated; the log(g) = 5 case suggests a mass of 30 M_Jup and age >10 Gyr.

Given that WISE 0647−6232 is in the far southern sky, could it be a member of one of the young kinematic groups primarily located in that region of the sky? Using our measured distance and transverse velocity (Table 3), we have analyzed whether there exists some value of the yet-unmeasured radial velocity that would yield U, V, W space motions suggesting membership in any of these young associations. We have checked published values of the U, V, W motions for the Argus/IC2391, TW Hydrae, Tucana/Horologium, β Pic, AB Dor, η Cha, Cha-Near, and Columba groups, and we find an excellent match to the Columba group if the unknown radial velocity of WISE 0647−6232 is later revealed to be ∼22 km s⁻¹. In fact, using the Bayesian technique advocated by Malo et al. (2013) we find a membership probability of 92.9% without assuming a value for the radial velocity, and this rises to 99.6% if the radial velocity is indeed 22 km s⁻¹. The age of 30 Myr for Columba would imply a mass of <2 M_Jup (Figure 2 of Burrows et al. 2003) for WISE 0647−6232 assuming an effective temperature range of 300–400 K. This, however, cannot be reconciled with the result from spectral model fitting (above) unless the true value of log(g) for WISE 0647−6232 is closer to 3.5 or 3.0.

We note, however, that the Y dwarf, at a distance of 8.7 pc, would be the closest member of Columba to the Sun; Torres et al. (2008) list the high probability members, and their distances range between 35 and 189 pc. If confirmed, WISE 0647−6232 would be the closest known member of the group; at a distance of 8.7 pc, it falls 2.4σ from the mean cluster distance of 82 ± 30 pc given by Torres et al. (2008). If the list of Columba members from Malo et al. (2013) is used instead, we find that the distance range is 35 to 81 pc. Similarly, the X, Y, Z values for WISE 0647−6232 are 2σ from the mean values noted in that paper. This indicates either a chance alignment with the space velocities of Columba or, if WISE 0647−6232 is a bona fide member, its low mass may have subjected it to kinematic heating and displaced it from the core of the moving group.

We thank our referee, Sandy Leggett, for a thoughtful and prompt report. We also thank Sergio Fajardo-Acosta for a careful reading of the manuscript. This publication makes use of data products from WISE, which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory (JPL)/California Institute of Technology (Caltech), funded by the National Aeronautics and Space Administration (NASA). This work is based in part on observations made with the Spitzer Space Telescope, which is operated by JPL/Caltech under a contract with NASA. Support for this work was provided by NASA through an award issued to programs 70062, 80109, and 90007 by JPL/Caltech. This work is also based in part on observations made with the NASA/ESA Hubble Space Telescope, obtained at STScI, which is operated by the Association of Universities for Research in Astronomy, Inc., under NASA contract NAS 5-26555. These observations are associated with programs 12970. Support for these programs was provided by NASA through a grant from STScI. This research has made use of the NASA/IPAC Infrared Science Archive, which is operated by JPL/Caltech, under contract with NASA. This paper includes data gathered with the 6.5 m Magellan Telescopes located at Las Campanas Observatory, Chile. C.G.T. acknowledges the support of ARC grants DP0774000 and DP130102695. Australian access to the Magellan Telescopes was supported through the National Collaborative Research Infrastructure Strategy of the Australian Federal Government. Travel support for Magellan observing was provided by the Australian Astronomical Observatory. This paper also includes Magellan data granted by the National Optical Astronomy Observatories (Proposal ID: 2010B-0184, P.I. Mainzer) through the Telescope System Instrumentation Program (TSIP). TSIP is funded by National Science Foundation. We are also indebted to the SIMBAD database, operated at CDS, Strasbourg, France.