Radial Velocities, Space Motions, and Nearby Young Moving Group Memberships of Eleven Candidate Young Brown Dwarfs
Adric R. Riedel, Victoria DiTomasso, Emily L. Rice, Munazza K. Alam, Ellianna Abrahams, James Crook, Kelle L. Cruz, Jacqueline K. Faherty
aa r X i v : . [ a s t r o - ph . S R ] A p r D RAFT VERSION A PRIL
25, 2019Typeset using L A TEX twocolumn style in AASTeX62
Radial Velocities, Space Motions, and Nearby Young Moving Group Memberships of Eleven Candidate Young Brown Dwarfs A DRIC
R. R
IEDEL ,
1, 2, 3, 4, 5 V ICTORIA D I T OMASSO ,
6, 4, 5 E MILY
L. R
ICE ,
7, 4, 8 M UNAZZA
K. A
LAM ,
9, 4, 5 E LLIANNA A BRAHAMS ,
10, 4, 11 J AMES C ROOK ,
4, 12, 13 K ELLE
L. C
RUZ ,
4, 5, 8
AND J ACQUELINE
K. F
AHERTY
4, 141
Space Telescope Science Institute, Baltimore, MD 21218, USA Department of Astronomy, California Institute of Technology, Pasadena, CA 91125, USA Department of Engineering Science and Physics, College of Staten Island, City University of New York, Staten Island, NY 10314, USA Department of Astrophysics, American Museum of Natural History, New York, NY 10024, USA Department of Physics and Astronomy, Hunter College, City University of New York, New York, NY 10065, USA Leibniz-Institute for Astrophysics Potsdam (AIP), An der Sternwarte 16, D-14482 Potsdam, Germany Department of Physics and Astronomy, College of Staten Island, City University of New York, Staten Island, NY 10314, USA Physics Program, The Graduate Center, City University of New York, 365 Fifth Ave, New York, NY 10016, USA Department of Astronomy, Harvard University, Cambridge, MA 02138, USA Department of Astronomy, University of California at Berkeley, Berkeley, CA 94720, USA Department of Physics, City College of New York, City University of New York, New York, NY 10031, USA Hunter College High School, 71 East 94th Street, New York, NY 10128, USA Physics and Astronomy Department, University of California Los Angeles, 430 Portola Plaza, Los Angeles, CA 90095, USA Department of Terrestrial Magnetism, Carnegie Institution of Washington, DC 20015, USA
Submitted to Astrophysical JournalABSTRACTWe present new radial velocity (RV) measurements for 11 candidate young very-low-mass stars and browndwarfs, with spectral types from M7 to L7. Candidate young objects were identified by features indica-tive of low surface gravity in their optical and/or near-infrared spectra. RV measurements are derived fromhigh resolution (R= λ / ∆ λ =20,000) J band spectra taken with NIRSPEC at the Keck Observatory. We com-bine RVs with proper motions and trigonometric distances to calculate three-dimensional space positions andmotions and to evaluate membership probabilities for nearby young moving groups (NYMGs). We propose2MASS J00452143+1634446 (L2 β , J =13.06) as an RV standard given the precision and stability of measure-ments from three different studies. We test the precision and accuracy of our RV measurements as a function ofspectral type of the comparison object, finding that RV results are essentially indistinguishable even with differ-ences of ± I lines at 1.24–1.25 µ mand evaluate their consistency with other age indicators. We confirm or re-confirm four brown dwarf membersof NYMGs – 2MASS J00452143+1634446, WISE J00470038+6803543, 2MASS J01174748 − − − Keywords: infrared: stars — techniques: radial velocities — stars: low-mass, brown dwarfs — techniques:spectroscopic INTRODUCTIONStudying brown dwarfs is our gateway to constrainingthe formation and evolutionary histories of giant planets Data presented herein were obtained at the W.M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology,the University of California and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support ofthe W.M. Keck Foundation. and their atmospheres. Brown dwarfs, especially youngobjects, can have masses and temperatures comparable todirectly-imaged exoplanets (Liu et al. 2013), but as free-floating objects rather than as stellar companions, they aremore amenable to detailed study with current instrumenta-tion. With the current generation of high contrast integralfield spectrograph instruments such as Project 1640, GPI, andSPHERE (Oppenheimer et al. 2013; Macintosh et al. 2008; R
IEDEL , D I T OMASSO ET AL .Beuzit et al. 2008) and soon JWST (Seager et al. 2009), thequestion of giant planet atmospheres and their formation isan increasing focus.Brown dwarfs do not achieve stable hydrogen fusion;therefore, they have no main sequence and no direct mass-luminosity relationship. Instead, brown dwarfs continuallydecrease in radius, temperature, and luminosity over time. Itis thus difficult to tell the difference between brown dwarfsof different masses based on spectra alone; a young low-mass brown dwarf can have the same temperature as an oldhigh-mass brown dwarf. Indeed, for many very-low-massobjects it is not possible to determine whether an object is astar or a brown dwarf without an estimate of the object’s age.There are two ways to resolve this mass-age degeneracy:dynamical mass measurement (e.g., Konopacky et al. 2010;Dupuy et al. 2014, 2015), which uses a combination of as-trometry and spectroscopy to determine dynamical masses;and age measurements, which currently rely on spectro-scopic and kinematic diagnostics. Dynamical masses requirethe brown dwarf to be in a close binary system, which israre (2.5 +8 . − . % of the population, Blake et al. 2010), and acomplete (or at least partial) orbit, which can require yearsto decades of astrometric monitoring. Precise age measure-ments for field-age and younger brown dwarfs (i.e., non-subdwarfs) require either a stellar companion with a reliableage constraint or membership in a nearby young movinggroup (NYMG), cluster, or star forming region where ageconstraints are then provided by the NYMG as a whole, typi-cally based on age constraints determined using higher-massmembers (e.g., Zuckerman & Song 2004).For young, single brown dwarfs, the most readily accessi-ble method to estimate age is via kinematic membership in aNYMG. The NYMGs are, as their name implies, groups ofstars and brown dwarfs moving together through space withsimilar space velocities. The assumption is that they formedtogether in a single star-forming event, with the same Galac-tic orbits as their natal molecular cloud. Though they arenot gravitationally bound to each other in an open cluster,they are still young enough to shear from the Galactic po-tential and that chance encounters with disk stars have notcompletely obscured their shared trajectory. As such, de-termining the space velocity (and space position) of youngobjects is a powerful method of determining their potentialmembership in a nearby young moving group. NYMGs aresparse, containing perhaps a few hundred members spreadout over thousands of cubic parsecs. Known groups in-clude β Pictoris ( ∼
20 Myr, Mamajek & Bell 2014), Tucana-Horologium ( ∼
45 Myr, Bell et al. 2015), Argus ( ∼
50 Myr,Barrado y Navascu´es et al. 1999) and AB Doradus ( ∼ > β Pictorisgiven only proper motion information; by that same token,the maximum probability rises to over 90% with the additionof radial velocity information, even without a distance. Ageconstraints provided by NYMG membership can range from5 Myr ( ǫ Chamæleon, Murphy et al. 2013) to 500 Myr ( χ For, P¨ohnl & Paunzen 2010) with uncertainties of ±
10 Myrfor TW Hydra (Weinberger et al. 2013) to ±
100 Myr forolder groups.The established memberships of NYMGs are deficient inlow-mass members (mid-M dwarfs and later) relative to thefield Initial Mass Function (e.g., Jeffries 2012; Kraus et al.2014; Gagn´e et al. 2017; Shkolnik et al. 2017). In order tocomplete the low-mass census of NYMGs, candidate young,very-low-mass objects are typically identified based on near-infrared (NIR) colors and low-resolution spectral features in-dicative of low surface gravity. Young very-low-mass objectsare typically 1-2 magnitudes redder than the average NIRcolor for their spectral type (Faherty et al. 2012). Spectra ofthese unusually red objects often exhibit spectroscopic signa-tures of low gravity, including weaker singly-ionized alkalimetal lines, which is often taken to be a sign of youth (e.g.,Cruz et al. 2009). These objects are assumed to be young,with spectral type suffixes coarsely defined according to thedivergence of gravity-sensitive spectral features from thoseof field (i.e., old) objects (Cruz et al. 2009; Allers & Liu2013). Finer age estimation based on spectral features aloneis not currently possible; therefore, establishing membershipin a NYMG is essential to providing age constraints for very-low-mass objects.There are currently over 160 objects with spectral typesM7 and later that have been identified as candidate mem-bers of nearby young moving groups. Prominent early ex-amples included TWA 27 (2MASS J12073346 − −
39) in TW Hydra (Gizis 2002), 2MASSJ01415823 − −
46, Kirkpatrick et al.2006) in Tucana-Horologium (Gagn´e et al. 2015a), 2MASSJ06085283 − −
27) in β Pic-toris (Rice et al. 2010, but see also Gagn´e et al. 2014cand Faherty et al. 2016), and 2MASS J03552337+1133437(hereafter 2M0355+11) in AB Doradus (Faherty et al. 2013;Liu et al. 2013). These objects have frequently been used ascomparison objects for newly discovered candidate younglow-mass objects and even directly-imaged exoplanets (e.g.,Miles et al. 2018; Greenbaum et al. 2018; Crepp et al. 2018).There are additionally over 150 very-low-mass stars andbrown dwarfs that display signatures of youth but lack
INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS K band (Blake et al. 2010) or the H band (Faherty et al.2016), we focus on the J band, which contains numerouswater absorption lines, prominent bandheads of FeH, andregions that are largely free of telluric absorption that canbe used for cross-correlated RV measurements (Prato et al.2015). The J band also contains strong alkali metal linesthat are sensitive to surface gravity (e.g., McLean et al. 2007;Rice et al. 2010). The objects in our sample could be verylow mass stars or brown dwarfs, depending on their ages, butwe refer to them as brown dwarfs for the sake of simplicity.In Section 2 we describe our sample of 11 M and L dwarfs,the NIRSPEC/Keck II observations, and the data reductionprocedure. In Section 3, we describe the analysis and re-sults, including RV measurements, the calculation of spacepositions and motions, and the five methods for estimatingNYMG membership probabilities. We present notes on re-sults for individual objects in Section 4. In Section 5, wediscuss implications of our results for measuring RV of late-type objects and for evaluating various youth indicators. Wepresent our conclusions in Section 6. SAMPLE, OBSERVATIONS, AND DATAREDUCTION2.1.
Sample Selection
Our targets were selected from a sample of ∼ M7 and laterdwarfs identified as candidate young objects via their theirclassification as low-surface-gravity objects by Cruz et al.(2009) using red-optical spectra and/or Allers & Liu (2013)using low-resolution, near-infrared spectra. All of the objectsalso have unusually red NIR colors for their spectral type (though not all in the specific J − W color shown in Fig-ure 1). Eleven objects were observed during four half-nightsin 2014 March and September; details of the observations aredescribed in Section 2.2 below and in Table 1.There are only minor discrepancies between the opticalspectra and the NIR spectroscopy seen here. The largest dif-ferences in spectral classification is 2MASS J02411151 − γ object by op-tical spectral typing (Faherty et al. 2016) but an L1 VL-G by infrared spectral typing (Allers & Liu 2013); andwith 2MASS J02535980+3206373 (hereafter 2M0253+32),which was a M7 β by optical typing (Faherty et al. 2016) butwas assigned as an M6 FLD-G in Allers & Liu (2013).All of these targets appear in Faherty et al. (2016), whereeight of them were identified as having kinematics thatsuggested possible membership in multiple NYMGs orthat could not be distinguished from field objects (“Am-biguous Member”). Two targets were determined byFaherty et al. (2016) to be bona fide group members:2MASS J00452143+1634446 (hereafter 2M0045+16), iden-tified as an Argus member by Gagn´e et al. 2014c, andWISE J004701.09+680352.2 (hereafter W0047+68), iden-tified as an AB Doradus member by Gizis et al. 2015and Liu et al. 2016. 2MASS 01174748 − −
34) was listed as a high-likelihood mem-ber of Tucana-Horologium by both Faherty et al. (2016)and Liu et al. (2016). Faherty et al. (2016) presentedRV measurements for three of our targets, one of which(2MASS J00452143+1634446, hereafter 2M0045+16) wasalso previously measured by Blake et al. (2010). The othertwo were measured from low-quality spectra, motivating ourdecision to observe them again. In the time between ourobservations and this publication, Gizis et al. 2015 reporteda radial velocity for W0047+68. These literature RV mea-surements are presented and compared to our RV results inSection 3.1. 2.2.
Observations
Observations were made UT 2014 May 22 & 24 andUT 2014 September 16 & 17 using NIRSPEC, the cryo-genic cross-dispersed echelle spectrometer on the Keck II10-m telescope at the W.M. Keck Observatory on Maunakea,Hawai’i (McLean et al. 1998, 2000). We used NIRSPEC’scross-dispersed echelle mode with the NIRSPEC-3 (N3) fil-ter, which approximates standard J band coverage (1.143–1.375 µ m). In echelle mode, eight usable dispersion orders(65 to 58) are captured on the detector. Because the spec-tral interval captured by the detector is slightly smaller thanthe free spectral range in each order, there are small gaps, in-creasing with wavelength, in the total spectral coverage. Theexact wavelength ranges for each dispersion order are listedin the headings in Section 2.4. The slit width is three pixels(0 . ′′ J band is approximately R= λ / ∆ λ =20,000 (“high” resolution)in echelle mode. Throughout the paper the high resolution J band spectra are referred to by the number of the dispersionorder, from 65 ( ∼ µ m) – 58 ( ∼ µ m). R IEDEL , D I T OMASSO ET AL . Figure 1.
Spectral type versus J − W color diagram for the11 sample objects (red points) plotted with the average J − W colors (black points) and 1 σ spreads (gray shaded region) fromFaherty et al. (2016). Eight of the 11 sample objects are more than1 σ redder in J − W color than the average for their spectral type,especially the L dwarfs. The remaining three are more than 1 σ redin other color combinations. Observing methods follow those described in detail byMcLean et al. (2007) and Prato et al. (2015); the followingis a brief summary and explanation of departures from thosemethods. Observations were made in pairs, nodding alongthe slit between each integration so that traces were sepa-rated by 7 ′′ on the 12 ′′ -long slit. Due to a desire to avoid anintermittent quadrant in the slit-viewing camera, recent highresolution observations have used a smaller nod length. Dur-ing these occasions the nod size was at least 2 ′′ so that thedispersed traces would be well-separated on the slit. Integra-tion time was 600 seconds per nod for all observations exceptfor all four exposures of 2M0253+32 and four (of eight to-tal) exposures of 2MASS J05341594-0631397 (2M0534-06),which were 480 seconds per nod.Total integration times per object are listed in Table 1.A0 V stars were observed at an airmass very close to thatof the target object (typically < Data Reduction
All of the observed data were reduced with the REDSPECIDL-based software package , described in McLean et al.(2003, 2007). The package performs standard bad pixel in-terpolation, dark subtraction, and flat-fielding as well as spa-tial rectification of curved spectra. Spectra are rectified and extracted in subtracted nod pairs so that the sky backgroundand OH emission lines are removed. Spectra were extractedby summing over 7–15 rows dependent on seeing, then sub-tracted again to produce a positive spectrum with residualsky emission features removed. Most orders were wave-length calibrated with OH night sky lines, which were foundto be highly stable and well-distributed across orders. Sevenhigh resolution dispersion orders were reduced, covering or-ders 58–65 with the exception of order 60, where the OHnight sky lines are blended with O emission bands at 1.26–1.28 µ m (Rousselot et al. 2000) making wavelength calibra-tion and sky subtraction considerably more difficult. Eachreduced spectrum was continuum normalized, and multiplenod pairs were averaged together to increase SNR. Spec-tra were shifted to the heliocentric reference frame usingbarycentric corrections calculated using JSkyCalc .2.4. Spectral Orders
We present all reduced NIRSPEC dispersion orders for2M0045+16 (L2 β , J =13.06) in Figure 2. We summarizethe relevant absorption features apparent in high resolutionM and L dwarf spectra by NIRSPEC dispersion order be-low. More details can be found in McLean et al. (2007) andRice et al. (2010). Order 58 (1.30447 – 1.32370 µ m) — Al I doublet in thecenter of the order. Order 59 (1.28262 – 1.30151 µ m) — Weak Fe I lines. Order 60 (1.26137 – 1.27999 µ m) — Not reduced (see 2.3above). Order 61 (1.24081 – 1.25913 µ m) — K I lines, higherSNR than order 65. Order 62 (1.22093 – 1.23899 µ m) — Used for RV mea-surements because of FeH and H O, strong well-spaced OHnight sky lines, and weak telluric lines.
Order 63 (1.20168 – 1.21938 µ m) — FeH and H O similarto order 62.
Order 64 (1.18293 – 1.20011 µ m) — Weak Ti I and Mn I lines. Order 65 (1.16496 – 1.18207 µ m) — K I lines, typicallylower SNR than in order 61 and blended with H O lines.Following the work of the NIRSPEC Brown Dwarf Spec-troscopic Survey (Prato et al. 2015) we use order 62 (1.221 –1.239 µ m) for our radial velocity measurements. This wave-length regime is essentially free of telluric features, and Mand L dwarf spectra contain numerous molecular absorption ∼ physics/labs/skycalc/flyer.html INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS Table 1.
Observing Log
Object 2MASS Optical NIR Sp. Type α δ J b Int. Time Average UT Date ofName a ID Sp. Type Sp. Type Ref J2000.0 J2000.0 mag seconds SNR Observation2M0253 +
32 02535980+3206373 M7 β M6 FLD-G 1,2 02 53 59.70 +
32 06 37.0 13.62 1920 21 2014 September 162M0534 −
06 05341594 − γ M8 VL-G 2,1 05 34 15.94 −
06 31 39.7 16.05 4320 5.4 2014 September 172M1935 −
28 19355595 − γ M9 VL-G 2,1 19 35 55.96 −
28 46 34.4 13.95 3600 25 2014 May 222M0027 +
05 00274197+0503417 M9.5 β L0 INT-G 2,1 00 27 41.97 +
05 03 41.7 16.19 4800 5.0 2014 September 162M0241 −
03 02411151 − γ L1 VL-G 3,1 02 41 11.50 −
03 26 58.0 15.80 4800 7.5 2014 September 162M0117 −
34 01174748 − β L1 INT-G 2,1 01 17 47.40 −
34 03 25.0 15.18 4800 10 2014 September 172M0045 +
16 00452143+1634446 L2 β L2 VL-G 3,1 00 45 21.43 +
16 34 44.6 13.06 1200 38 2014 September 162M1551 +
09 15515237+0941148 L4 γ L4 VL-G 2,1 15 51 52.37 +
09 41 14.8 16.32 7200 7.3 2014 May 242M1615 +
49 16154255+4953211 L4 γ L3 VL-G 2,1 16 15 42.50 +
49 53 21.0 16.79 7200 4.5 2014 May 222M2154 −
10 21543454 − β L5 γ −
10 55 30.0 16.44 5400 4.3 2014 May 24W0047 +
68 00470038+6803543 L7 ( γ ?) L7.5 pec 2,5 00 47 01.06 +
68 03 52.1 15.60 4800 6.9 2014 September 17 a b From 2MASS All-Sky Point Source Catalog.
References —Optical spectral types are those defined in Cruz et al. (2009), and near-infrared spectral types are on the scale defined in Allers & Liu (2013). Spectral typesuffixes indicate the strength of gravity-sensitive features, with β is roughly equivalent to INT - G and γ to VL - G . Individual references are: (1) Allers & Liu (2013), (2)Faherty et al. (2016), (3) Cruz et al. (2009), (4) Gagn´e et al. (2014b), (5) Gizis et al. (2012). lines from FeH and H O that are ideal for cross-correlationtechniques. Figure 3 presents the order 62 spectra for all 11objects in the sample. Our targets have brightnesses between J =13 and J =17. Even with total integration times of 20 min-utes to two hours (listed in Table 1), the resulting spectrahave average SNR in order 62 between 4 and 38, with a max-imum SNR=38.2 for 2M0045+16 ( J =13.06) and minimumSNR=4.3 for 2M2154 −
10 ( J =16.44).We also tested RV measurements using order 59, which isfree of strong telluric absorption and was used by Prato et al.(2015) for cross-correlating spectra of T dwarfs. For M andL dwarfs, the intrinsic atomic lines and molecular absorptionlines at these wavelengths are weaker and the results were farless reliable. No order 59 results are presented in this paper. ANALYSIS3.1.
Radial Velocity Measurements
To measure radial velocity (RV), we use a customcross-correlation code written in Python, first described inFaherty et al. (2016). The inputs are a heliocentric-correctedstellar spectrum (wavelength, flux, and uncertainty) and thespectrum of a comparison object with a previously measuredRV, taken with the same instrumental setup to avoid system-atics. For comparison spectra, we use objects with NIRSPECBrown Dwarf Spectroscopic Survey (BDSS) spectra fromMcLean et al. (2007), Rice et al. (2010), and Prato et al.(2015). For the radial velocities of these standards, we usevalues reported in Rice et al. (2010), Blake et al. (2010),Chubak et al. (2012), and Prato et al. (2015), listed in Table2. The target and standard spectra are read in and interpolatedonto a log-normal spaced wavelength grid covering only theregion where the spectra overlap, up-sampled in wavelengthby a factor of 10. A third-order fit to the spectra is removed,
Table 2.
Comparison Stars
Name Spectral RV RV AverageType km s − Ref. SNRGJ 406 M6 19.321 ± − ± −
20 M6.5 − ± − ± − ± ± − ± −
27 M8.5 23 ± ± ± a L0.5 52.37 ± ± −
02 L1.5 − ± − ± −
10 L2.5 − ± β − ± ± −
01 L4.5 − ± −
16 L5 − ± a Binary
References —(1)Chubak et al. (2012), (2) Rice et al. (2010), (3) Prato et al.(2015), (4) Blake et al. (2010) taking out the large-scale structure of the spectra and leav-ing only the spectral lines. The code then re-samples eachflux point of the target and comparison spectra from withinthe estimated noise on each, which we model as Gaussianrandom noise. The resulting re-sampled target spectrum is R
IEDEL , D I T OMASSO ET AL . Figure 2.
NIRSPEC spectra for dispersion orders 58, 59, and 61 to65 for the L2 object 2M0045+16. cross-correlated with the re-sampled radial velocity compar-ison spectrum, and the cross-correlation results are fit witha Gaussian+linear function to determine the velocity shift inpixels. The process of re-sampling the noise is repeated 1000times, producing 1000 velocity shift measurements for eachtarget and comparison object spectra pair, which we bin intoa histogram. The histogram of velocity measurements is fitwith a Gaussian function, the mean of which we adopt as thevelocity shift, and the 1- σ width of which we adopt as theuncertainty on the measurement, accounting for the noise ofboth spectra. The result then is converted from pixel shiftsto velocity in km s − , and the known velocity of the com-parison object is then subtracted to provide the actual helio-centric radial velocity of the target. The final uncertainty ofeach RV measurement is the combination of the uncertaintyfrom the cross-correlation procedure and the uncertainty inthe previously measured RV of the comparison object, addedin quadrature. The latter dominates the precision of the re-sults. Figure 3.
NIRSPEC dispersion order 62 spectra (1.22093 – 1.23899 µ m) for 11 objects in the sample. This cross-correlation technique is subject to comparisonobject-dependent systematic errors, including uncertaintieson the wavelength solution of each observation and on thepreviously measured radial velocity of the comparison ob-ject. To test the accuracy of our RV results, we cross-correlate our targets with all 19 comparison objects in Ta-ble 2 using spectra from Prato et al. (2015) and Rice et al.(2010), producing 19 individual RV measurements for eachof our target objects. The mean, weighted by the calculateduncertainty, of these individual RV results constitutes the fi-nal calculated RV for each target object, as listed in Table 3.This procedure was used to calculate the RVs of eight out ofeleven of our target objects.The remaining three objects’ RVs were calculated withslight modification to the routine described above. For thesethree targets, the cross-correlation process produced outly-ing pixel shifts that implied unrealistic velocities. For thosethree objects, W0047+68, 2M0534 −
06, and 2M0241 − ±
500 or ±
100 re-sampled wavelength pixels, de-
INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS −
03, six of the 19 comparison objects pro-duced outlying pixel shifts. For two comparison objects,SCH J1612 −
20 and DENIS-P J1605 −
24, restricting the ac-ceptable pixel shifts to ranges of ±
50 or fewer re-sampledwavelength pixels still failed to produce a Gaussian distri-bution of measured RV-induced pixel shifts, so we omittedthese two comparison objects in the calculation of the finalRV for 2M0241 −
03. For the other four comparison objects,it took an average of 1087 iterations to produce 1000 ac-ceptable pixel shifts; the maximum number of iterations was1206 and the minimum was 1008.For 2M0534 −
06, all of the comparison spectra producedoutlying pixel shifts, but with restricting the allowed pixelshift produced Gaussian results for all comparison spectra. Ittook an average of 6236 iterations to produce 1000 accept-able pixel shifts; the maximum was 59794 and the minimumwas 1015. 3.2.
Space Positions and Motions
As yet, no single photometric or spectroscopic youth indi-cator can be used to assign a precise and reliable age estimatefor low mass stars and brown dwarfs. Thus, we rely on kine-matics – positions in RA and DEC (hereafter α , δ , and π );motions in µ R.A. cos decl , µ decl (hereafter µ α ∗ , µ δ and RV)– to determine if the brown dwarfs are likely members of anearby young moving group (NYMG). Even so, kinematicsare necessary but not sufficient to prove youth. The sheernumber of disk stars and the large kinematic space occupiedby these unbound groups mean field-age disk stars can be in-terlopers, hence the importance of spectroscopic indicationsthat the brown dwarfs are in fact young.There are two basic strategies for determining membershipin moving groups. One is to take the three positional ob-servables ( α , δ , and π ) and three velocity observables ( µ α ∗ , µ δ , RV), convert them into three-dimensional cartesian spacepositions (XYZ, where X is toward Galactic center, Y isin the direction of solar motion, and Z is toward the NorthGalactic Pole) and three-dimensional cartesian space veloci-ties (UVW, where U is motion along the X axis, V along Y,and W along the Z axis), and compare the star’s UVWXYZvalues to those of the moving group, represented as ellipsoidsfrom Riedel et al. 2017 in Figures 4 and 5. The other methodis to reverse the process: take the UVWXYZ properties of themoving group, translate them to observable quantities like µ ,RV, and π , at the α and δ of the target star, and compare thepredicted values of a group member to the actual values ofthe target star.It is possible to handle incomplete kinematic data. Indeed,one object in this study lacks a parallax measurement. Ifconverting to UVWXYZ, a range of reasonable parallax val- ues can be tested to see if any are consistent with movinggroup membership. If converting to observables, it is possi-ble to simply not run a comparison against the predicted par-allax value. This ensures that we can still evaluate kinematicmemberships, though at the cost of reduced membership cer-tainty. For a more complete discussion of the dependence ofmembership probabilities on observed data, see Riedel et al.(2017)Positions for our targets come from the 2MASS (Cutri et al.2003) catalog, with the sole exception of W0047+68from WISE (Cutri & et al. 2012). Proper motions weresourced from a variety of papers, principally Gagn´e et al.(2014c), Faherty et al. (2016), Gaia Collaboration et al.(2018), Casewell et al. (2008), and other papers listed inTable 3. All of our targets have more than one proper motionmeasurement. Most of the proper motions are relative mea-surements from catalog surveys or parallax programs andare consistent at the 1- σ level for a given object, though allare generally in agreement to within ±
10 mas yr − , withthe exception of 2M0241 −
03, one uncertain measurement of2M0117 −
34, and one extremely uncertain measurement of2M1615+49. We list them all individually in Table 3.We use published parallax measurements from Faherty et al.(2016), Dieterich et al. (2014), Zapatero Osorio et al. (2014),Dahn et al. (2002), Liu et al. (2016), Gizis et al. (2015), andGaia Collaboration et al. (2018) to obtain complete spacemotions (Table 3, Figures 4 and 5) and more confident groupmembership probabilities. Seven objects have multiple par-allax measurements, which are often discrepant from eachother by more than 1- σ . In the case of 2M0241 −
03, thethree parallaxes are only consistent at the 2- σ level, andin the case of 2M0045+16, the Gaia Collaboration et al.(2018) parallax is consistent with the Liu et al. (2016)parallax but not the Zapatero Osorio et al. (2014) paral-lax. In the case of 2M0253+32, the Faherty et al. (2016)parallax implies a larger distance than Liu et al. (2016)or Gaia Collaboration et al. (2018), though neither dis-tance makes the target a more likely member of anyknown NYMG. 2M1935 −
28 is the only case where theGaia Collaboration et al. (2018) and Liu et al. (2016) par-allaxes do not agree even at the 2- σ level, though both dis-tances independently make the brown dwarf a β Pic member.With 2M0027+05, the Liu et al. (2016) parallax disagreeswith Dahn et al. (2002), though neither parallax produces alikelihood of membership in any known NYMG.Radial velocities have already been published for 2M0045+16(Blake et al. 2010; Faherty et al. 2016), W0047+68 (Gizis et al.2015), 2M0241 −
03 (Faherty et al. 2016), 2M1615+49(Faherty et al. 2016), and 2M1935 −
28 (Shkolnik et al.2017). Our result for 2M0045+16 is consistent with thosemeasured by Blake et al. (2010) and Faherty et al. (2016)well within the 1- σ uncertainties, even of the most precisemeasurement ( ± − ). Our result for W0047+68 isconsistent with the previous measurement to within ∼ σ ofour lower-precision measurement. Both of the Faherty et al.(2016) measurements were relatively low precision ( ± − ), upon which our measurements improve by about R IEDEL , D I T OMASSO ET AL .a factor of two or more, and our results are consistent towithin 1- σ uncertainties.All of the assembled measurements were combined withstandard weighted means and weighted standard deviations.All individual results, and the weighted values (shown inbold) actually used in membership probability analysis, areshown in Table 3.3.3. Membership Probabilities
There are a number of differing approaches to kinematicmoving group identification, and following Faherty et al.(2016), we consider the results of five different codes (thefour used in Faherty et al. 2016 for comparison purposes,plus the newer BANYAN Σ code) to provide consensus ap-proach. Four of the codes used here (BANYAN I, BANYANII, BANYAN Σ , and LACEwING, see descriptions and ref-erences below), test against different properties: proper mo-tion, radial velocity, parallax, and space position. The fifthcode, the convergence code from Rodriguez et al. (2013),considers only a single test for proper motion, but predictsthe distance and radial velocity. If data do not exist or arenot present, the test is simply not run and the codes produceresults based only on the other tests.The Convergence code presented in Rodriguez et al.(2013) exploits the fact that if all of the stars in a movinggroup exhibit parallel space motions, their proper motionvectors should converge at a point in space (a ”convergentpoint”, analogous to the vanishing point) because of perspec-tive effects. The code computes probabilities of membershipin six moving groups (TW Hya, β Pic, Tuc-Hor, Columba,Carina-Near, and AB Dor) by comparing the proper mo-tion vector defined by ( µ α ∗ , µ δ ) to one pointing toward theconvergent point of a given moving group. From there it pre-dicts the associated radial velocity and distance of the object,which can be compared to any actual measurements.BANYAN I (Malo et al. 2013) uses a Bayesian formal-ism to evaluate which of seven nearby young moving groups(TW Hya, β Pic, Tuc-Hor, Columba, Carina, Argus, andAB Dor) or a field population of which an object is mostlikely to be a member. It converts observables to Cartesianspace. Unlike the Convergence code, radial velocity and par-allax measurements are incorporated into the probability dis-tribution rather than predicted according to possible groupmembership.BANYAN II (Gagn´e et al. 2014c) is a modification ofBANYAN I. It considers the same seven moving groups asBANYAN I, but it is based on a different set of bona-fidemembers, assumes an uneven distribution of the number ofstars in each group, and allows freely oriented moving groupsin space. It takes a hybrid approach, constraining observablesbased on Cartesian space.LACEwING (Riedel et al. 2017) predicts memberships in16 nearby young moving groups and open clusters within 100pc: compared to the BANYAN codes, it adds ǫ Chamæleon, η Chamæleon, 32 Orionis, Octans, Carina-Near, ComaBerenices, Ursa Major, χ Fornax, and the Hyades. LikeBANYAN II, all groups are represented as freely-oriented ellipsoids with numerically proportionate populations. Un-like BANYAN II, it does not use Bayesian priors; insteadit relies on the parameterized results of a large simulationof stars to translate goodness-of-fit values into membershipprobabilities. It operates in observational space.BANYAN Σ (Gagn´e et al. 2018) is a more refined ver-sion of BANYAN II using multivariate Gaussian models(instead of the orthogonal axis models of BANYAN II andLACEwING) which predicts memberships in 27 youngmoving groups and associations with ages up to 800 Myrand distances up to 150 pc, including all of the groups inLACEwING, plus 118 Tau, Corona Australis, Upper CoronaAustralis, IC 2391, IC 2602, Lower Centaurus Crux, UpperCentaurus Lupus, Upper Sco, ρ Oph, the Pleiades, Taurus,Platais 8, Volans-Carina, and the new formulation of Ar-gus identified in Zuckerman (2019). Like BANYAN II, itconstrains the observables based on Cartesian space. MEMBERSHIP RESULTS AND NOTES ONINDIVIDUAL OBJECTSWe identify five of our 11 sample objects as high prob-ability members of known NYMGs. Four of these arere-confirmations of possible memberships presented inGizis et al. (2015), Faherty et al. (2016), Liu et al. (2016),or Shkolnik et al. (2017), and one is a new membership. Theremaining six objects are found to have no membership in aknown NYMG. — (L2 β ) was identified by Gagn´e et al. (2014b)as a member of the roughly 50 Myr old Argus associa-tion, using more or less the original definition of Argusfrom Torres et al. (2008) and by Liu et al. (2016) usingBANYAN II. That identification is reconfirmed here with90-100% probabilities, which maintains this object as oneof the few brown dwarfs in Argus, with an estimated massof 25.0 ± jup (Faherty et al. 2016). Given that Argushas been kinematically (Torres et al. 2008) and chemically(De Silva et al. 2013) associated with the nearby IC 2391open cluster, we can draw on the properties of hundredsof higher mass stars to understand this and other similarlyyoung brown dwarfs. 2M0045+16 is a member of Argus,both in its original formulation (Torres et al. 2008 and subse-quent, used in BANYAN I, BANYAN II, and LACEwING)which was disputed by Bell et al. (2015), and the new defini-tion from Zuckerman (2019) (used in BANYAN Σ ).There are three published parallaxes for 2M0045+16, twoof which agree with each other, while a third value fromZapatero Osorio et al. (2014) is inconsistent at a 2- σ level(see Table 3). Even using discrepant parallax, we find Ar-gus to be the most likely NYMG membership by all methodsthat consider membership in Argus.As alternative hypothesis, LACEwING suggests that2M0045+16’s kinematics are also consistent with β Pictoris.This would make 2M0045+16 a significantly younger browndwarf of roughly 25 Myr (Mamajek & Bell 2014) rather than50 Myr.
INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS Figure 4.
Projected UVW space motions in the UV, UW, and VW planes for the objects from our sample with apparent memberships in aknown NYMG. The black ellipse denotes the UVW phase-space position of the object relative to the known NYMGs and nearby open clusters(taken from Riedel et al. 2017), which are shown with 1- σ extents in different colors. IEDEL , D I T OMASSO ET AL . Figure 5.
Same as Figure 4, showing the five objects from our sample with parallaxes but no apparent membership in a known NYMG. The Vvelocity of 2M0253+32 is − km s − , which is outside the range of our plots. INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS J =13.06 mag makes it the 6th brightest L2and in the top 25 brightest early ( < L5) L dwarfs, just 0.6 magfainter than the brightest known L2 and ∼ K -band observations (3.29 ± − ) and Faherty et al.(2016) from 2008 H -band observations (3.16 ± − ),both also from NIRSPEC. Given the consistency of these andour measurement of 3.29 ± J -bandobservations, it seems that 2MASS J0045+16 is RV stableand an optimal late-type spectral standard. W0047+68 — (L7 γ ) has previously been identified as anAB Doradus member by Gizis et al. (2015) and Liu et al.(2016) using full UVWXYZ space motion and positionfitting. We reconfirm that membership: W0047+68 is amember of AB Doradus according to every code, despitea 2 σ disagreement between our radial velocity and that ofGizis et al. (2015). This L7 γ object is one of the least mas-sive known free-floating extrasolar objects, with an estimatedmass of 11.8 ± Jup (Faherty et al. 2016), despite beingsubstantially older than other brown dwarfs with a γ gravityclassification. − — (L1 β ) is confirmed with our RV measurementand the parallax from Gaia Collaboration et al. (2018) as amember of Tucana-Horologium, an identification made byFaherty et al. (2016) solely on the basis of its proper mo-tion and by Liu et al. (2016) on the basis of its proper mo-tion and parallax. The new membership is agreed uponby every moving group code and implies that the browndwarf is 16.4 ± Jup (Faherty et al. 2016). The alterna-tive proper motions from Casewell et al. (2008), Gagn´e et al.(2014c) and Liu et al. (2016) have much larger motion alongthe α axis than the proper motion calculated by Faherty et al.(2016), but all the membership codes still find membershipin Tucana-Horologium. − — (M9 γ ) is a member of β Pictoris, first iden-tified as such by Shkolnik et al. (2017). With LACEwING,BANYAN I, and BANYAN Σ , it is a moderate or high prob-ability member; with BANYAN II, it is lower likelihood;and with the Convergence method, it is either a β Pictorisor Columba member (see Table 5). We consider this systemas a high probability member of β Pictoris. − — (L4 β ) is identified by LACEwING, the Con-vergence Code, and BANYAN Σ as moderate probabilitymember of Carina-Near, a 200 Myr old group identified byZuckerman et al. (2006). This makes 2M2154 −
10 the old-est confirmed NYMG member in the sample. Gagn´e et al.(2014c) found 2M2154 −
10 to be a member of Argus, whichwe do not reproduce due to a disagreement in the RV: as amember of Argus it should have an RV of roughly −
14 kms − , while we measure an RV of − ± − . The remaining targets — (2M0027+05 [M9.5 β ], 2M0253+32[M7 β ], 2M0534 −
06 [M8 γ ], 2M0241 −
03 [L0 γ ], 2M1551+09[L4 γ ], and 2M1615+49 [L4 γ ] were all identified havingambiguous NYMG membership by Faherty et al. (2016)and are not found to be likely members of any knownNYMG with the addition of our RV measurements andGaia Collaboration et al. (2018) astrometry.The convergence method predicts that 2M0253+32 is amember of β Pic with a predicted RV of 5.5 km s − , butour measured RV is − − . We therefore concludethat 2M0253+32 is not a β Pic member.2M1615+49 appears to be a rapid rotator (see Figure 10)but does not otherwise distinguish itself. LACEwING findsit to be a potential member of AB Doradus, though at lowprobability; the Convergence code finds it a possible mem-ber of Tuc-Hor (with a predicted RV of −
15 km s − , whichdoes not match our measured −
24 km s − ), and BANYAN I,BANYAN II, and BANYAN Σ find it is not a member of anygroup at a probability above the threshold of interest.The expected RV for 2M1551+09 if it were a memberof β Pic, −
17 km s − , is consistent with the actual mea-sured velocity of −
15 km s − , but only the convergencecode finds that membership and at a low probability. Theexpected distance for a β Pic member with the proper mo-tion of 2M1551+09 would place it very far spatially fromthe known members of β Pic (a condition the BANYANcodes and LACEwING consider), which means that even ifits parallax-determined distance matches the expected dis-tance of 30 pc, the object cannot be a member.2M0241 −
03 has five published proper motions (Faherty et al.2016; Zapatero Osorio et al. 2014; Gagn´e et al. 2014c;Casewell et al. 2008; Liu et al. 2016) and three paral-laxes (Faherty et al. 2016; Zapatero Osorio et al. 2014;Liu et al. 2016), which only agree with each other at the2 σ level. This system has been considered a member ofTucana-Horologium since Gagn´e et al. (2014c), but with ourweighted parallax we find no such membership. Using theFaherty et al. (2016) and Zapatero Osorio et al. (2014) par-allaxes individually, the brown dwarf is still not a memberof any moving group. Liu et al. (2016) placed it in Tuc-Horwith an 82% likelihood using BANYAN II (a lower proba-bility than our threshold for BANYAN II) using parallax andproper motion, but with our astrometry BANYAN II gives usan 88% membership (also below the threshold) in β Pic in-stead. LACEwING does reproduce membership in Tuc-Horat a low 30% probability, and we note that LACEwING givesa higher (46%) chance of membership in Columba. TheConvergence code suggests a low probability of membershipin Carina-Near if the system is at 80 pc, which it is not.Ultimately, the reason it is not in Tuc-Hor is a combinationof factors: if its (combined) proper motion and radial veloc-ity were to imply the best possible space velocity match toTuc-Hor, the brown dwarf would need to be closer to 64 pcaway, which not even the Liu et al. (2016) parallax (54 pc)agrees with, while (simultaneously) being at that appropriatedistance would put it approximately 40 pc away (over 2- σ )from the bulk of the Tuc-Hor moving group.2 R IEDEL , D I T OMASSO ET AL .For now, we suspect that these objects are members of ayoung field population, which Riedel et al. (2017) has shownto be quite substantial. DISCUSSION5.1.
RV Measurements of Very-Low-Mass Objects
Typically, high-resolution spectra are cross-correlatedagainst spectra of objects with similar spectral types, ef-fectively doubling the required observing time, which canbe on the order of several hours for intrinsically faint very-low-mass stars and brown dwarfs. For example, Prato et al.(2015) used a similar cross-correlation method to ours formeasuring the radial velocities of very-low-mass objects, butrestricted that comparison to objects with similar spectraltypes. In order to optimize the efficiency of RV measure-ments for very-low-mass objects, we test the dependence ofthe precision and accuracy of RV results on the spectral typeof the comparison object.Figure 6 shows the spectral type of each RV comparison,as listed in Table 2, as a function of the calculated RV ofthe target object. The gray bar indicates the 1- σ uncertaintyon the final RV measurement of the target, as listed in Ta-ble 3. The three panels show the objects of earliest and lat-est spectral type in our sample, 2M0253+32 (M7, top panel)and W0047+68 (L7, bottom panel), as well as the object ofmedian spectral type, 2M0045+16 (L2, middle panel). Thesame test was done for all eleven objects in our sample. Wesee no correlation between spectral type of the comparisonobject and the precision or accuracy of the individual RVmeasurement, compared to the final value. Thus, we showthat a cross-correlation comparison objects can be as differ-ent as ± NIR Colors of Young Very-Low-Mass Objects
By virtue of their selection as objects with spectral indica-tors of youth, most of our sample has redder near- and mid-infrared colors than expected for normal L dwarfs (Figure 1),with the exception of 2M0253+32, which is consistent withthe colors of a normal L dwarf of the same spectral type.The degree of reddening is fairly consistent across all near-and mid-infrared color combinations, with a few exceptions:In H − K , 2M0027+05 is bluer than (but consistent with) anormal L dwarf; in K − W , 2M0534-06 is likewise blueror consistent with a normal L dwarf. Neither of these effects Figure 6.
Measured RV (km s − ) versus spectral type of each com-parison object for 2M0253+32 (top), 2M0045+16 (middle), andW0047+68 (bottom). The gray bar represents the RV 1- σ uncer-tainty range on the final results for each target. The RV of the targetis calculated by cross-correlating each target and comparison pair,and does not show any correlation to or dependence on the differ-ence in spectral type between the comparison and target objects. INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS Wavelength ( m)
Wavelength ( m)
Figure 7. K I triplet for 2M0253+32 (left panel) and 2M1935 − are due to poor precision photometry; they appear to be real(or perhaps variable) features of the objects themselves.The NIR colors alone are not a sufficient gauge of age. Themost consistently discrepant objects in the sample, whichare also generally the objects with the largest color offsetsfrom normal brown dwarfs, are 2M0241 −
03, 2M1615+49,2M2154 −
10, and W0047+68; W0047+68 is identified as anAB Dor (120 Myr) member and is both the potentially oldestidentified member in the sample and the coldest brown dwarfin the sample, while the newly-identified β Pictoris (25 Myr)member, 2M1935 −
28, is just above the envelope of youngbrown dwarfs in Figure 1.5.3. K I Line Strengths
Measurement of gravity-sensitive spectral lines may pro-vide a more reliable indicator of youth than NIR colors alone(e.g., Faherty et al. 2012). The neutral alkali metal absorp-tion lines like those of Na I and K I are weaker in lower sur-face gravity atmospheres (e.g., Schlieder et al. 2012), whichtranslates to smaller equivalent widths (EWs). This providesa way to test if red objects are truly young and low surfacegravity, or simply red because of dustier atmospheres. Withour high-resolution spectra we can measure the strength andwidth of gravity-sensitive lines and test this directly, whichwe describe below. We can also determine if these objectsare rapid rotators by measuring the Full Width at Half Maxi-mum (FWHM) of the lines.NIRSPEC orders 61 (1.24–1.26 µ m) and 65 (1.165–1.182 µ m) each contain a K I doublet (in order 65, thelines are sometimes resolvable into a triplet, see Fig-ure 7) that has been shown to be sensitive to tempera-ture (e.g., McLean et al. 2007) and surface gravity (e.g.McGovern et al. 2004; Rice et al. 2010). Therefore we mea-sure these line strengths and compare them to those of field-age objects to evaluate additional indicators of youth for oursample. Because of the only occasionally resolved triplet inorder 65, which is also typically of lower SNR, we concen-trate our analysis on the order 61 doublet.Following the methods of Alam & Douglas (2016), wequantify the strengths of the ∼ µ m K I lines for our sam-ple by computing EWs and FWHM using PHEW: PytHon Wavelength (µm) N o r m a li z e d F l u x Figure 8.
Subsection of NIRSPEC dispersion order 61 for2M0045+16 centered on the λ =1.2525 K I line to demonstratethe line strength measurement methods of Alam & Douglas (2016).The yellow line represents the defined pseudocontinuum, the purplecurve the Voigt profile, and the purple horizontal line the FWHM.The shaded green region represents the equivalent width. Equivalent Widths code. We measure EWs using th -orderfit to the pseudo-continuum, defined as the average flux out-side of the absorption line within a 1.241–1.246 µ m window,and a Voigt profile fit to the absorption line. The equivalentwidth is calculated by integrating the pseudo-continuum levelminus the spectrum over the selected range. Uncertaintieswere estimated via 1,000 Monte Carlo iterations. We reportthe means and standard deviations of these measurements inTable 5. Figure 8 presents an example of these measurementsfor the λ =1.2525 K I line from the spectrum of 2M0045+16.Line strength measurements, compared with results for fieldobjects from Alam et. al (in prep.) and McLean et al. (2003),are presented in Figures 9 and 10. The complete dataset forboth field and suspected young objects is presented in Table5. These results follow the general pattern indicated by otherstudies of the K I lines, e.g. Allers & Liu (2013) (Figure23), Gagn´e et al. (2015b) (Figure 6), and Martin et al. (2017)(Figure 3) and indicate that our suspected young sample ex-hibits lower surface gravities, as expected for objects with β and γ gravity designations. Our results are not directlycomparable to those of the aforementioned papers due toour higher spectral resolution. Those papers used moderate(R ∼ ∼ µ m line and correspondingly more distinction be-tween field objects and our suspected young sample, even https://github.com/munazzaalam/PHEW/ IEDEL , D I T OMASSO ET AL . M6 M8 L0 L2 L4 L6 L8Optical Spectral Type0246810121416 . μ m E W ( Å ) FieldβγM6 M8 L0 L2 L4 L6 L8Optical Spectral Type0.02.55.07.510.012.515.017.5 . μ m E W ( Å ) Fieldβγ
Figure 9. K I equivalent width versus optical spectral type forour sample of candidate young, unusually red objects organized bygravity class (red stars, blue circles) compared to field objects (graytriangles) for the order 61 lines at 1.2436 µ m (top panel) and 1.2525 µ m (bottom panel) lines. at the extremes where Allers & Liu (2013) could only deter-mine an EW-based gravity classification and saw no differ-ences between young and field stars in their K I index.The FWHM measurements (Figure 10) demonstrate thatalmost all of our targets have lower v sin i than the fieldobjects. Two objects, 2M0241 −
03 (L0 γ ) and 2M1615+49(L4 γ ), are possibly rapid rotators and/or viewed more edge-on than the other young objects, which would broaden theirgravity-weakened lines, as evidenced by their higher FWHMmeasurements but similar EWs to the other candidate youngobjects. 5.4. Consistency of Age Indicators
Near-infrared colors may indicate possible youth, but theyhave never been considered sufficient to determine spe-cific ages for young brown dwarfs, as noted in Section 5.2.Spectroscopic measurements are more useful for evaluatingyouth, but here too there are limitations. All of our objectshave been classified as either β or γ gravity classes accord- M6 M8 L0 L2 L4 L6 L8Optical Spectral Type50100150200250300350 . μ m F W H M ( k m / s ) Fieldβγ
Figure 10. K I FWHM versus optical spectral type for our sampleof candidate young, unusually red objects compared to field objectsfor the order 61 line at 1.2525 µ m line. Colors are the same asFigure 9. ing to their red-optical spectra (Table 1), and most objectshave INT-G or VL-G gravity classes on the near-infraredAllers & Liu (2013) spectral system. As shown in Figure 9,all of our targets have weaker (lower EW) K I doublet linesthan field-age dwarfs of comparable spectral type, indicatingyouth, though there is substantial overlap between β and γ gravity classifications. Furthermore, gravity-related spectraltype suffixes themselves do not appear to track directly withage; our 50 Myr old Argus member 2M0045+16 is an L2 β ,while our 125 Myr old AB Doradus member W0047+68 isclassified as L7 γ . We therefore cannot assign even relativeages based on line strength or gravity measurements alone.We also do not consider it problematic that W0047+68 showssigns of youth when M dwarf members of AB Doradus typ-ically do not have identifiable low surface gravity features(Schlieder et al. 2012), as it is a much cooler, thus lower-mass, object that may be evolving more slowly.Even the non-spectrophotometric property of kinematicmemberships has limitations. Taken in total, the linestrengths and colors indicate that all the objects in our sam-ple are young (if not precisely how young), but we can onlyconnect five of them with NYMGs that confirm a young age.The failure to connect the remaining objects to a NYMGcould be explained by one of four possibilities:1. The NYMG identification algorithm may be flawed be-cause it is based on inaccurate or incomplete assump-tions of how to best identify NYMGs.2. We may have insufficiently precise kinematic data forthe late-type object, or an inaccurate understanding ofthe parameters of the NYMG itself.3. The object may be a member of an as-yet-unknownNYMG. INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS CONCLUSIONSIn this paper we have presented new high-resolution NIRspectroscopy of 11 red, low-gravity, late-type objects. Us-ing new RV measurements derived from that spectroscopy,and proper motion and parallax measurements from litera-ture sources, we re-confirm membership of four objects inNYMGs. We also identify a new member of Carina-Nearand confidently rule out six objects as members of the knownNYMGs. These objects remain interesting targets for study,though we cannot currently determine their ages or origins.Our study also adds more evidence to the hypothesis (suchas proposed by Riedel et al. 2017 and Gagn´e et al. 2018) thatthere are other populations of young objects in the solarneighborhood yet to be discovered, whether they are newNYMGs or a genuinely unassociated ”field” population ofyoung objects. The six objects we conclusively rule out asmembers of the known NYMGs are an indistinguishable pop-ulation, spectroscopically and photometrically, from the con-firmed NYMG members. There is no reason to say that theyare not young, beyond lack of group membership.We also presented evidence that the accuracy of the cross-correlation technique is not dependent on close spectral typematches. Previously, it was thought that spectral types ofstandard stars had to be as close as possible to the spec-tral type of the target - within two subtypes - for the cross-correlation radial velocity technique. The power of this tech-nique in face of the spectral type discrepancy is due to thestrength and regularity of the FeH lines in cool star spectra,the rectification step where a third-order polynomial fit corre-sponding to the overall shape of the spectrum is removed, andthe use of multiple comparison spectra in a weighted mea-surement. The end result is proof that collecting an extensive library of standards at every spectral type is not necessaryto achieve kilometer-per-second precision radial velocities,and therefore shows that the technique of cross-correlation ischeaper and easier to implement than previously thought.The authors wish to thank the staff of the Keck Obser-vatory for their outstanding support, including Luca Rizzi,Jim Lyke, Cynthia Wilburn, Terry Stickel, Jason McIlroy,Heather Hershley, and Barbara Schaefer. Observing assis-tance from Kay Hiranaka was greatly appreciated. The au-thors are grateful for assistance with references from J.T.Wright and E.R. Newton via Twitter, and J. Gagne for helpwith the BANYAN II and BANYAN Σ codes. ARR, VD,and ELR were responsible for writing the majority of the pa-per. ARR was responsible for the youth analysis, conclu-sions, and system notes. VD and EA were responsible forthe NIRSPEC data reduction and description in the paper.VD, ELR, EA, and ARR were responsible for the RV anal-ysis. MKA was responsible for the line measurements, anddescription of PHEW. KLC was responsible for the spectraltyping and advising of VD and EA. JKF was responsible forthe general survey outline.This material is based upon work supported by theNational Science Foundation under grant numbers AST-1313278, AST-1313132, and AST-1153335. EA acknowl-edges support from the National Science Foundation Grad-uate Research Fellowship under Grant No. DGE 1752814.This research was supported in part by NASA through theAmerican Astronomical Society’s Small Research Grant Pro-gram. This research has made use of the NASA/IPAC In-frared Science Archive, which is operated by the Jet Propul-sion Laboratory, California Institute of Technology, undercontract with the National Aeronautics and Space Admin-istration. This publication makes use of data from the TwoMicron All Sky Survey, which is a joint project of the Uni-versity of Massachusetts and the Infrared Processing andAnalysis Center, funded by the National Aeronautics andSpace Administration and the National Science Foundation.This research has benefited from the M, L, and T dwarf com-pendium housed at DwarfArchives.org and maintained byChris Gelino, Davy Kirkpatrick, and Adam Burgasser. Thisresearch has made use of the SIMBAD database, operatedat CDS, Strasbourg, France and NASA’s Astrophysics DataSystem. his work was supported by a NASA Keck PI DataAward, administered by the NASA Exoplanet Science Insti-tute. Keck telescope time was granted by NOAO, through theTelescope System Instrumentation Program (TSIP). TSIP isfunded by NSF. Data presented herein were obtained at theW. M. Keck Observatory from telescope time allocated to theNational Aeronautics and Space Administration through theagency’s scientific partnership with the California Institute ofTechnology and the University of California. The Observa-tory was made possible by the generous financial support ofthe W. M. Keck Foundation. The authors wish to recognizeand acknowledge the very significant cultural role and rev-erence that the summit of Maunakea has always had within6 R IEDEL , D I T OMASSO ET AL .the indigenous Hawaiian community. We are most fortunateto have the opportunity to conduct observations from thismountain.
Facility:
Keck:II (NIRSPEC)REFERENCES
Alam, M., & Douglas, S. 2016, PHEW: PytHon EquivalentWidths, , , doi:10.5281/zenodo.47889.https://doi.org/10.5281/zenodo.47889Allers, K. N., & Liu, M. C. 2013, ApJ, 772, 79Barrado y Navascu´es, D., Stauffer, J. R., Song, I., & Caillault, J.1999, ApJL, 520, L123Bell, C. P. M., Mamajek, E. E., & Naylor, T. 2015, MNRAS, 454,593Beuzit, J., Feldt, M., Dohlen, K., et al. 2008, in Society ofPhoto-Optical Instrumentation Engineers (SPIE) ConferenceSeries, Vol. 7014, Society of Photo-Optical InstrumentationEngineers (SPIE) Conference SeriesBinks, A. S., & Jeffries, R. D. 2014, MNRAS, 438, L11Blake, C. H., Charbonneau, D., & White, R. J. 2010, ApJ, 723, 684Casewell, S. L., Jameson, R. F., & Burleigh, M. R. 2008, MNRAS,390, 1517Chubak, C., Marcy, G., Fischer, D. A., et al. 2012, ArXiv e-prints,arXiv:1207.6212Crepp, J. R., Principe, D. A., Wolff, S., et al. 2018, ApJ, 853, 192Cruz, K. L., Kirkpatrick, J. D., & Burgasser, A. J. 2009, AJ, 137,3345Cutri, R. M., & et al. 2012, VizieR Online Data Catalog, 2311Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, 2MASS AllSky Catalog of point sources.Dahn, C. C., Harris, H. C., Vrba, F. J., et al. 2002, AJ, 124, 1170De Silva, G. M., D’Orazi, V., Melo, C., et al. 2013, MNRAS, 431,1005Dieterich, S. B., Henry, T. J., Jao, W.-C., et al. 2014, AJ, 147, 94Dupuy, T. J., Liu, M. C., & Ireland, M. J. 2014, ApJ, 790, 133Dupuy, T. J., Liu, M. C., & Leggett, S. K. 2015, ArXiv e-prints,arXiv:1502.04707Faherty, J. K., Burgasser, A. J., Cruz, K. L., et al. 2009, AJ, 137, 1Faherty, J. K., Rice, E. L., Cruz, K. L., Mamajek, E. E., & N´u˜nez,A. 2013, AJ, 145, 2Faherty, J. K., Burgasser, A. J., Walter, F. M., et al. 2012, ApJ, 752,56Faherty, J. K., Riedel, A. R., Cruz, K. L., et al. 2016, ApJGagn´e, J., Burgasser, A. J., Faherty, J. K., et al. 2015a, ApJL, 808,L20Gagn´e, J., Faherty, J. K., Cruz, K., et al. 2014a, ApJL, 785, L14Gagn´e, J., Lafreni`ere, D., Doyon, R., et al. 2014b, ApJL, 792, L17Gagn´e, J., Lafreni`ere, D., Doyon, R., Malo, L., & Artigau, ´E.2014c, ApJ, 783, 121—. 2015b, ApJ, 798, 73Gagn´e, J., Faherty, J. K., Cruz, K. L., et al. 2015c, ApJS, 219, 33 Gagn´e, J., Faherty, J. K., Mamajek, E. E., et al. 2017, ApJS, 228,18Gagn´e, J., Mamajek, E. E., Malo, L., et al. 2018, The AstrophysicalJournal, 856, 23. http://stacks.iop.org/0004-637X/856/i=1/a=23Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018,ArXiv e-prints, arXiv:1804.09365Gizis, J. E. 2002, ApJ, 575, 484Gizis, J. E., Allers, K. N., Liu, M. C., et al. 2015, ApJ, 799, 203Gizis, J. E., Faherty, J. K., Liu, M. C., et al. 2012, AJ, 144, 94Greenbaum, A. Z., Pueyo, L., Ruffio, J.-B., et al. 2018, AJ, 155,226Jameson, R. F., Casewell, S. L., Bannister, N. P., et al. 2008,MNRAS, 384, 1399Jeffries, R. D. 2012, in EAS Publications Series, Vol. 57, EASPublications Series, ed. C. Reyl´e, C. Charbonnel, &M. Schultheis, 45–89Kirkpatrick, J. D., Barman, T. S., Burgasser, A. J., et al. 2006, ApJ,639, 1120Konopacky, Q. M., Ghez, A. M., Barman, T. S., et al. 2010, ApJ,711, 1087Kraus, A. L., Shkolnik, E. L., Allers, K. N., & Liu, M. C. 2014, AJ,147, 146Lew, B. W. P., Apai, D., Zhou, Y., et al. 2016, ApJL, 829, L32Liu, M. C., Dupuy, T. J., & Allers, K. N. 2016, ApJ, 833, 96Liu, M. C., Magnier, E. A., Deacon, N. R., et al. 2013, ApJL, 777,L20Macintosh, B. A., Graham, J. R., Palmer, D. W., et al. 2008, inSociety of Photo-Optical Instrumentation Engineers (SPIE)Conference Series, Vol. 7015, Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference SeriesMalo, L., Doyon, R., Lafreni`ere, D., et al. 2013, ApJ, 762, 88Mamajek, E. E., & Bell, C. P. M. 2014, MNRAS, 445, 2169Marocco, F., Andrei, A. H., Smart, R. L., et al. 2013, AJ, 146, 161Martin, E. C., Mace, G. N., McLean, I. S., et al. 2017, ApJ, 838, 73McGovern, M. R., Kirkpatrick, J. D., McLean, I. S., et al. 2004,ApJ, 600, 1020McLean, I. S., Graham, J. R., Becklin, E. E., et al. 2000, in Societyof Photo-Optical Instrumentation Engineers (SPIE) ConferenceSeries, Vol. 4008, Society of Photo-Optical InstrumentationEngineers (SPIE) Conference Series, ed. M. Iye & A. F.Moorwood, 1048–1055McLean, I. S., McGovern, M. R., Burgasser, A. J., et al. 2003, ApJ,596, 561McLean, I. S., Prato, L., McGovern, M. R., et al. 2007, ApJ, 658,1217
INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS McLean, I. S., Becklin, E. E., Bendiksen, O., et al. 1998, in Societyof Photo-Optical Instrumentation Engineers (SPIE) ConferenceSeries, Vol. 3354, Society of Photo-Optical InstrumentationEngineers (SPIE) Conference Series, ed. A. M. Fowler, 566–578Miles, B. E., Skemer, A. J., Barman, T. S., Allers, K. N., & Stone,J. M. 2018, ApJ, 869, 18Murphy, S. J., Lawson, W. A., & Bessell, M. S. 2013, MNRAS,435, 1325Newton, E. R., Charbonneau, D., Irwin, J., et al. 2014, AJ, 147, 20Oppenheimer, B. R., Baranec, C., Beichman, C., et al. 2013, ApJ,768, 24P¨ohnl, H., & Paunzen, E. 2010, A&A, 514, A81Prato, L., Mace, G. N., Rice, E. L., et al. 2015, ApJ, 808, 12Rice, E. L., Barman, T., Mclean, I. S., Prato, L., & Kirkpatrick,J. D. 2010, ApJS, 186, 63Riedel, A. R., Blunt, S. C., Lambrides, E. L., et al. 2017, AJ, 153,95Rodriguez, D. R., Zuckerman, B., Kastner, J. H., et al. 2013, ApJ,774, 101 Rousselot, P., Lidman, C., Cuby, J.-G., Moreels, G., & Monnet, G.2000, A&A, 354, 1134Schlieder, J. E., L´epine, S., Rice, E., et al. 2012, AJ, 143, 114Schmidt, S. J., West, A. A., Hawley, S. L., & Pineda, J. S. 2010,AJ, 139, 1808Seager, S., Deming, D., & Valenti, J. A. 2009, Astrophysics andSpace Science Proceedings, 10, 123Shkolnik, E. L., Allers, K. N., Kraus, A. L., Liu, M. C., & Flagg, L.2017, AJ, 154, 69Thompson, M. A., Kirkpatrick, J. D., Mace, G. N., et al. 2013,PASP, 125, 809Torres, C. A. O., Quast, G. R., Melo, C. H. F., & Sterzik, M. F.2008, Young Nearby Loose Associations, ed. B. Reipurth (SanFrancisco, Astronomy Society of the Pacific), 757–+Weinberger, A. J., Anglada-Escud´e, G., & Boss, A. P. 2013, ApJ,762, 118Zapatero Osorio, M. R., B´ejar, V. J. S., Miles-P´aez, P. A., et al.2014, A&A, 568, A6Zuckerman, B. 2019, ApJ, 870, 27Zuckerman, B., Bessell, M. S., Song, I., & Kim, S. 2006, ApJ, 649,L115Zuckerman, B., & Song, I. 2004, ARA&A, 42, 685
IEDEL , D I T OMASSO ET AL . INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS Table 3.
Astrometric and RV Data
Name α δ
Ref. µ α ∗ σ µα ∗ µ δ σ µδ Ref. π σ π Ref. RV σ RV Ref. ◦ ◦ mas yr − mas yr − mas yr − mas yr − mas mas km s − km s − − − − − − − −
06 83.566445 − − −
28 293.983154 − − − − − − − − − − − −
03 40.297996 − − − − − − − −
34 19.447838 − −
45 8 (9) 25.56 0.71 (8) 3.258 1.351 (3)111.5 2.1 − − − − − −
40 10 (4) 57.3 2 (4) 3.287 1.333 (3)354.4 2.2 − −
26 12 (7) 65.02 0.23 (8) 3.16 0.83 (9)374.9 8.5 − − − −
70 22 −
50 22 (15) 22.1 1.5 (5) − − − − − − − − −
80 12 18 12 (9) 32.0 1.0 (5) − −
23 34 41.8 45.5 (16) − − − − − − − − − W0047+68 11.751611 68.065102 (11) 387 4 −
197 4 (12) 82 3 (12) − −
210 10 (13) 82.3 1.8 (5) − − − − − N OTE —Data sources: (1) Cutri et al. (2003) [2MASS], (2) Dahn et al. (2002), (3) This Work, (4) Zapatero Osorio et al. (2014), (5) Liu et al. (2016), (6) Blake et al. (2010), (7)Jameson et al. (2008), (8) Gaia Collaboration et al. (2018), (9) Faherty et al. (2016), (10) Gagn´e et al. (2014c), (11) Cutri & et al. (2012) [WISE], (12) Gizis et al. (2015), (13)Thompson et al. (2013), (14) Casewell et al. (2008), (15) Faherty et al. (2009), (16) Schmidt et al. (2010), (17) Gagn´e et al. (2014a), (18) Gizis et al. (2012), (19) Shkolnik et al.(2017). Values in bold are weighted means.
IEDEL , D I T OMASSO ET AL . Table 4. Spatial & Kinematic Properties
Object Sp. Type X Y Z U V WName (Optical) pc pc pc km s − km s − km s − β − ± ± − ± ± − ± ± −
28 M9 β ± ± − ± − ± − ± − ± β − ± ± − ± − ± ± − ± −
03 L0 γ − ± ± − ± − ± − ± − ± −
34 L1 β − ± − ± − ± − ± − ± ± γ − ± ± − ± − ± − ± − ± −
10 L4 β ± ± − ± − ± − ± ± γ ± ± ± − ± − ± − ± γ ± ± ± − ± − ± − ± γ − ± ± ± − ± − ± − ± INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS
212 R
IEDEL , D I T OMASSO ET AL . Table 5. Membership Results
Object Sp. Type Final LACEwING BANYAN I BANYAN II CONVERGE a BANYAN Σ Name (Optical) Membership2M0253 +
32 M7 β None None Field Field ( β Pic-92) a Field2M0534 −
06 M8 γ None None Field Argus-94 (AB Dor-100) Field2M1935 −
28 M9 γ β
Pic β Pic-63 β Pic-100 β Pic-100 β Pic-98 β Pic-992M0027 +
05 M9.5 β None None Field Field (Car-Near-100) Field2M0241 −
03 L0 γ None Field Field Field Field Field2M0117 −
34 L1 β Tuc-Hor Tuc-Hor-96 Tuc-Hor-100 Tuc-Hor-100 Tuc-Hor-99 Tuc-Hor-1002M0045 +
16 L2 β Argus Argus-98 Argus-100 Argus-100 Field b Argus-1002M1551 +
09 L4 γ None None Field Field Field Field2M1615 +
49 L4 γ None AB Dor-25 Field (AB Dor-37) (Tuc-Hor-86) Field2M2154 −
10 L4 β Carina-Near Carina-Near-53 Field b Field Field Car-Near-89W0047 +
68 L7 γ AB Dor AB Dor-100 AB Dor-100 AB Dor-100 AB Dor-85 AB Dor-100N
OTE —The quoted membership probability is the highest membership probability for the most commonly identified mov-ing group, considering every permutation of kinematic data. Probabilities in parentheses are below the quality threshold:LACEwING: 20%; BANYAN codes, 90%; Convergence code, 80%. See discussion in Section 3.3. a Values in parentheses are inconsistent with actual membership. Either the probability is too low for that particular code,or (particularly for the Convergence code) the predicted distance, space position, or radial velocities are inconsistent withmembership or actual measurements. b The Convergence Code and BANYAN Σ do not consider membership in Argus. Only LACEwING, the Convergence code, andBANYAN Σ consider membership in Carina-Near. INEMATIC P ROPERTIES OF Y OUNG B ROWN D WARFS Table 6.
Spectral Line Measurements µ m 1.2525 µ mObject Name Optical SpT EW ( ˚A) FWHM (km/s) SNR EW ( ˚A) FWHM (km/s) SNRSuspected Young Objects2M0253+32 M7 β ± ± ± ± γ ± ± ± ± −
28 M9 γ ± ± ± ± β ± ± ± ± −
03 L0 γ ± ± ± ± −
34 L1 β ± ± ± ± β ± ± ± ± −
10 L4 β ± ± ± ± γ ± ± ± ± γ ± ± ± ± γ ± ± ± ± a M6 5.60 ± ± ± ± a M9 9 ± ± ± ± a L0 11.50 ± ± ± ± a L0.5 14.1 ± ± ± ± b L1 10.1 ± ± ± ± b L1 10.22 ± ± ± ± − b L1.5 10.91 ± ± ± ± − b L1.5 11.88 ± ± ± ± a L2 14.10 ± ± ± ± − b L2.5 11.05 ± ± ± ± b L3 11.34 ± ± ± ± b L3.5 11.40 ± ± ± ± a L4 14.00 ± ± ± ± b L4.5 9.69 ± ± ± ± − b L4.5 12.81 ± ± ± ± b L5 4.18 ± ± ± ± − a L5 14.70 ± ±
27 1.00 10.00 ± ±
24 1.002MASS J0103+1935 b L6 7.46 ± ± ± ± − b L6 2.60 ± ± ± ± − a L7 8.20 ± ± ± ± OTE —Objects are grouped by spectral type and then listed in order of right ascension. a Originally published in McLean et al. (2007). bb