DISCOVERY OF FOUR HIGH PROPER MOTION L DWARFS, INCLUDING A 10 pc L DWARF AT THE L/T TRANSITION^*,†

W1405+8350 was observed on UT Date 2012 March 23 with the 1.55 m Strand Astrometric Reflector at the Flagstaff Station of the US Naval Observatory (USNO) using a Tek2K 2048 × 2048 CCD Camera. We obtained I_C and z band images of W1405+8350. The data were reduced using standard techniques and calibrated to the known Sloan Digital Sky Survey (SDSS) magnitude (z = 15.28 AB mag; Aihara et al. 2011) of SDSS J152702.74+434517.2, observed immediately beforehand. The observed magnitudes of W1405+8350 are z = 17.50 ± 0.04 and I_C = 18.93 ± 0.06.

We used similar criteria to search for high proper motion objects as Gizis et al. (2011b), but extended the search to red colors. We searched for WISE sources that had detections at W1 (3.4 μm), W2 (4.6 μm), and W3 (12 μm), and no 2MASS counterpart within 3''. Sources in the WISE catalog are already matched to 2MASS sources within 3''. By selecting sources without a 2MASS counterpart, with a difference in all-sky surveys of about a decade, this constrained the search to sources with a proper motion of >03 yr⁻¹. WISE sources meeting our criteria were examined visually using 2MASS and WISE finder charts⁸ in order to look for apparent motion of a source between the two surveys. The discoveries presented here are only a subset of the entire survey, the full catalog of high proper motion objects will be reported in a future paper (J. E. Gizis & P. J. Castro, in preparation). WISE J040137.21+284951.7, WISE J040418.01+412735.6, and WISE J062442.37+662625.6 were discovered from the WISE preliminary data release⁹ and are reanalyzed using the WISE all-sky data release,¹⁰ and WISE J140533.32+835030.5 was discovered from the WISE all-sky data release.

WISE J140533.32+835030.5 (W1405+8350) was found to have a separation of ≈85 from a 2MASS source to the southeast, 2MASS J14053729+8350248. 2MASS J14053729+8350248 was previously overlooked due to its proximity (≈11'') to a bright background star, 2MASS J14053168+8350188 (J = 10.71 and V = 11.75; Lasker et al. 2008), resulting in a photometric confusion flag for the J band ("cc_flg" of "c00"). The object would otherwise have been identified in previous searches of 2MASS for ultracool dwarfs (Kirkpatrick et al. 2000; Cruz et al. 2003). In our USNO z band imaging data, J140532.57+835031.7 was detected ≈2'' to the northwest of W1405+8350, along the line of motion between the 2MASS and WISE sources, confirming W1405+8350 as a high proper motion object. The WISE source shows colors that are red, W1 − W2 = 0.58 ± 0.04, consistent with that of a late L dwarf/early T dwarf (Kirkpatrick et al. 2011); the 2MASS source has red colors, J − K_s = 1.78 ± 0.06, that are consistent with an L dwarf (Kirkpatrick et al. 2000); and the USNO source has red colors, I_C − z = 1.43 ± 0.07, consistent with a late L dwarf (Dahn et al. 2002). We positively identify the 2MASS and USNO sources as W1405+8350 at their respective epochs. With a high proper motion indicating a nearby object and red colors in 2MASS, WISE, and USNO indicating a late spectral type, we claim the detection of a nearby ultracool dwarf. A finder chart for W1405+8350 showing a clear linear sequence of positions at the epoch of 2MASS, WISE, and USNO is shown in Figure 1.

Zoom In Zoom Out Reset image size

WISE J040137.21+284951.7 (W0401+2849) was found to have a separation of ≈6'' from a 2MASS source to the east, 2MASS J04013766+2849529. W0401+2849 is detected on several of the Palomar Observatory Sky Survey (POSS) plates, POSS2 red and POSS2 IR, with at best a mediocre detection on POSS1 red, confirming the apparent motion between 2MASS and WISE. See Figure 2 for the finder chart. The WISE source shows colors that are red, W1 − W2 = 0.27 ± 0.03, consistent with that of a late M dwarf/early L dwarf (Kirkpatrick et al. 2011); the 2MASS source has red colors, J − K_s = 1.59 ± 0.03, that are consistent with an early L dwarf (Kirkpatrick et al. 2000). This source was just outside of the J band magnitude cut (<20 pc) in a search for ultracool dwarfs in the 2MASS Second Incremental Release by Cruz et al. (2003).

Zoom In Zoom Out Reset image size

WISE J040418.01+412735.6 (W0404+4127) was found to have a separation of ≈4'' from a 2MASS source to the north, 2MASS J04041807+4127398. W0404+4127 is detected on several of the POSS plates, POSS2 red and POSS2 IR, confirming the apparent motion between 2MASS and WISE. It is however too faint to be detected on the POSS1 red plate. See Figure 2 for the finder chart. The WISE source shows colors that are red, W1 − W2 = 0.30 ± 0.03, consistent with that of a late M dwarf/early L dwarf (Kirkpatrick et al. 2011); the 2MASS source has red colors, J − K_s = 1.73 ± 0.04, that are consistent with an L dwarf (Kirkpatrick et al. 2000). This source was likely overlooked because it lies at a low galactic latitude, b = −8°, where previous searches often avoided the galactic plane (Gizis et al. 2000; Cruz et al. 2003).

WISE J062442.37+662625.6 (W0624+6626) was found to have a separation of ≈65 from a 2MASS source to the northwest, 2MASS J06244172+6626309. 2MASS J06244172+6626309 is reported as a source having poor photometry with 2MASS "PH_QUAL"¹¹ flags of "EEA" for the JHK_s bands, respectively. W0624+6626 was originally reported as an extended source in the WISE preliminary data release, WISEP J062442.27+662626.1, with the all-sky release resolving it into two distinct sources, W0624+6626 and WISE J062441.53+662629.2. WISE J062441.53+662629.2 lies on the position of the 2MASS source to the northwest, 2MASS J06244172+6626309, with W1 − W2 ≈ 0, it is likely earlier than M0 (Kirkpatrick et al. 2011). WISE J062441.53+662629.2 appears to be a background star that W0624+6626 was in close proximity to during the 2MASS epoch, resulting in poor 2MASS photometry; likely this is the reason W0624+6626 was previously overlooked. W0624+6626 is detected on several of the POSS plates, POSS1 red, POSS2 red, and POSS2 IR, confirming that it passed very close to a background star during the 2MASS epoch; see Figure 2. W0624+6626 shows colors that are red, W1 − W2 = 0.24 ± 0.04, consistent with that of a late M dwarf/early L dwarf (Kirkpatrick et al. 2011); while the 2MASS J and H band photometry are contaminated by a background star.

To determine the proper motion of W1405+8350 we used the background stars within 6' of the WISE source of interest as reference to calibrate the positions on the USNO z band image in the 2MASS astrometric reference frame. We find a proper motion directly from the USNO z and 2MASS astrometry for W1405+8350 to be μ_αcos (δ) = −0.63 ± 001 yr⁻¹ and μ_δ = 0.57 ± 002 yr⁻¹, with total motion 0.85 ± 002 yr⁻¹.

We determine the proper motions of W0401+2849 and W0404+4127 between the WISE and 2MASS epochs relative to the common background stars within 5' of the WISE source of interest. We find a proper motion for W0401+2849 of μ_αcos (δ) = −0.48 ± 001 yr⁻¹ and μ_δ = −0.10 ± 001 yr⁻¹, with total motion 0.49 ± 002 yr⁻¹. We find a proper motion for W0404+4127 of μ_αcos (δ) = −0.06 ± 001 yr⁻¹ and μ_δ = −0.40 ± 001 yr⁻¹, with total motion 0.40 ± 001 yr⁻¹.

Since W0624+6626 is coincident with a background star for the 2MASS epoch, the proper motion needs to be determined using epochs in which W0624+6626 is resolved from the background star. The only two epochs where this is the case is the WISE epoch and POSS1 red epoch; see Figure 2. We determine the proper motion by using the background stars within a 5' × 5' field of view centered on the WISE source of interest as reference to calibrate the positions on the POSS1 red image in the WISE astrometric reference frame. We find a proper motion directly from the WISE and POSS1 red astrometry for W0624+6626 to be μ_αcos (δ) = 0.355 ± 0002 yr⁻¹ and μ_δ = −0.513 ± 0003 yr⁻¹, with total motion 0.624 ± 0004 yr⁻¹.

A transformation between I_C and i is determined by using data from Table 2 of Dahn et al. (2002) and SDSS DR8 photometry. We selected L dwarfs from Dahn et al. (2002) that have I_C photometry and are not reported as photometrically variable or spectroscopically peculiar. For those sources that had counterparts in SDSS DR8 (Aihara et al. 2011), we obtained i and z band photometry, resulting in 12 sources with I_C, i, and z band photometry. One source was removed, the L4.5 dwarf 2M2224-01, due to conflicting z band photometry from Dahn et al. (2002) and SDSS, Δz ≈ 0.7 mag, well outside of the uncertainty, suggesting erroneous data or variability. Last, the matching between Dahn et al. (2002) and SDSS DR8 was verified visually. We plot i–z versus I_C–i and find the best linear least-squares fit to the data; see Figure 3. The black dashed line shows the best fit line and the red dotted lines show the 1σ error in the fit:

$$\begin{eqnarray*} I_{\rm C}-i &=& (-0.0953 \pm -0.0647) + (-0.5010\, {\pm} \, 0.0349) (i-z) \\ &&\quad {\rm for}\ 1.7 \lesssim i-z \lesssim 3.1. \end{eqnarray*}$$

Zoom In Zoom Out Reset image size

Using this transformation, we find a value of i = 20.56 ± 0.16 for W1405+8350 from I_C and z photometry, where the error is from the uncertainty in the photometry and the transformation. Additional observations of close (bright) mid- to late L dwarfs within the coverage of SDSS in the I_C band would further constrain the fit. The blue dash-dot line is the transformation from Lupton (2005) shown for comparison, which overlaps our transformation and its uncertainty for mid-L dwarfs (L5–L7), but strongly deviates for early L dwarfs and to a smaller degree the latest L dwarfs.

Our transformation was originally used to estimate the optical spectral type of W1405+8350, by comparing W1405+8350 to the SDSS L dwarfs from Schmidt et al. (2010), using color–color diagrams in the same manner as Castro & Gizis (2012). Our transformation from I_C to i for L dwarfs will be useful to the astronomical community, enabling the direct comparison of L dwarfs with I_C and z photometry to the large catalog of L dwarfs from SDSS (Schmidt et al. 2010).

Moderate-resolution spectra of W1405+8350 were obtained with the Near-Infrared Spectrometer (NIRSPEC; McLean et al. 1998, 2000) at the 10 m W. M. Keck Observatory on 2012 June 7 (UT) and processed using the IDL software package REDSPEC (McLean et al. 2003). Spectra were obtained in the N3 configuration with data spanning the J band, 1.14 to 1.36 μm. We used the 42 × 057 slit, giving a resolving power of R ∼ 2000. Seeing was ≈13. Two ABBA sequences of 120 s were taken, with a total integration time of 960 s. The standard HD172864 (A2V) was observed immediately after the science observation at similar airmass for telluric correction. Wavelength calibration made use of the Ne and Ar calibration lamps. The spectrum has been normalized at 1.27 μm and corrected for heliocentric motion.

W1405+8350 was also observed using TripleSpec (Wilson et al. 2004) at the Apache Point Observatory (APO) 3.5 m telescope on UT Date 2012 June 2 under non-photometric conditions. The spectrum was extracted and calibrated using a custom version of SpeXTool (Vacca et al. 2003; Cushing et al. 2004). The A-star calibrator was HD 99966. The resulting spectrum is low signal-to-noise ratio.

Optical spectra were obtained on UT Dates 2012 August 26 and 27 using the MMT Red Channel Spectrograph (RCS; NOAO Proposal ID: 2012B-0233). Spectra were obtained with the 10 slit aligned at the parallactic angle. Grating 270 (blazed at 7300 Å) with order-blocking filter LP-530 was used to yield wavelength coverage 6200–9800 Å and resolution 12 Å; strong telluric water absorption limited the usefulness of the data beyond 9000 Å. Conditions were non-photometric. The data were extracted and calibrated using standard IRAF tasks.

Low-resolution NASA Infrared Telescope Facility (IRTF) SpeX (Rayner et al. 2003) spectra were obtained on 2012 February 15 (UT) and processed using SpeXTool (Vacca et al. 2003; Cushing et al. 2004). Spectra were obtained in prism mode using the 05 slit aligned at the parallactic angle. The resolution of the corresponding data spanning 0.7–2.5 μm was λ/Δλ ≈ 120. Conditions were non-photometric with light clouds and the seeing was >1'' at K. We obtained individual exposures for each object in an ABBA dither pattern along the slit. Immediately after the science observations we observed an A0V star for each object at a similar airmass for telluric corrections and flux calibration. Internal flat field and Ar arc lamp exposures were acquired for pixel response and wavelength calibration, respectively.

4.1.1. Near-infrared and Optical Spectroscopy

Figure 4 shows the moderate-resolution NIRSPEC near-infrared (NIR) J band spectrum of W1405+8350 (black) compared to NIR standards (red) and a reference dwarf (blue) from the Brown Dwarf Spectroscopy Survey (BDSS)¹² (McLean et al. 2003) and the moderate-resolution SpeX library¹³ (Cushing et al. 2005). While the formal classification standards are also reference dwarfs, we identify this subset of reference dwarfs separately since they are our primary reference of comparison for spectroscopic classification. We use the NIR standards of Kirkpatrick et al. (2010) as anchors to determine the spectral classification. The K i doublet lines at 1.168, 1.177 μm and 1.243, 1.254 μm are sensitive to surface gravity and thus an indicator of age (McLean et al. 2003; McGovern et al. 2004). The equivalent widths of the K i doublets increase as a function of spectral type from mid-M dwarfs to mid-L dwarfs, decreasing for late L dwarfs, and increasing again for the early T dwarfs; with the late L dwarfs showing significant scatter, especially the L8 dwarfs. There are FeH bands from 1.1939 to 1.2389 μm, present in mid-M dwarfs, strengthening until about L5, then decaying and are absent in the T dwarfs. A doublet of Al i at 1.311 and 1.314 μm is present in mid-M dwarfs but vanishes near the M/L boundary (McLean et al. 2003). The absence of the Al i doublet indicates the spectral type is later than M, while the presence of the shallow FeH bands indicate W1405+8350 must be a late L dwarf or a very early T dwarf. Note the deeper FeH bands for the L_n7 standard compared to that of the shallower features of W1405+8350. The deeper K i doublets of W1405+8350 compared to the L8/L9 dwarfs indicate it is an old field dwarf. From the four deep absorption features of K i we find a radial velocity of −27 ± 14 km s⁻¹. We use the goodness-of-fit χ² as defined in Burgasser (2007b) to compare W1405+8350 to the NIR templates; from 1.15 to 1.33 μm, omitting regions with the deep K i absorption lines (1.165 to 1.185 μm and 1.24 to 1.26 μm, since they are gravity sensitive) and >1.33 μm due to a deviation in wavelength calibration for W1405+8350 relative to the templates. The best fit to W1405+8350 is the L_n9 standard DENIS-P J0255−4700, we adopt an NIR spectral type of L9 for W1405+8350.

Zoom In Zoom Out Reset image size

Figure 5 shows the smoothed TripleSpec NIR spectrum of W1405+8350 (black) compared to NIR standards (blue, red) (Kirkpatrick et al. 2010; Burgasser et al. 2006a) observed with SpeX¹⁴ at low-resolution. The TripleSpec spectrum is low signal-to-noise, with the H and K bands being shown and the J band omitted due to significant noise; the smoothed noise (gray) is shown at the bottom. The overall appearance of the H band and K band regions is consistent with the L9 NIR spectral classification.

Zoom In Zoom Out Reset image size

Figure 6 shows the optical spectrum for W1405+8350 without telluric corrections. Optical spectral types for the L dwarfs were determined by comparing the appearance of the features over the range 6500 Å to 9000 Å with the standards defined by Kirkpatrick et al. (1999). The best fit for W1405+8350 is the L8 optical standard 2MASSW J1632291+190441, we adopt an optical spectral type of L8 for W1405+8350.

Zoom In Zoom Out Reset image size

4.1.2. Distance and Physical Properties

We estimate the distance by using the spectral-type–absolute-magnitude relationships of Looper et al. (2008a) for 2MASS photometry and the spectral-type–absolute-magnitude relationships of Dupuy & Liu (2012) for 2MASS and WISE photometry. We find a distance of 9.5 ± 1.3 pc from 2MASS J photometry, 9.7 ± 1.3 pc from 2MASS H photometry, and 10.0 ± 1.5 pc from 2MASS K_s photometry using the relations from Looper et al. (2008a), 9.8 ± 1.8 pc from 2MASS J photometry, 9.6 ± 1.8 pc from 2MASS H photometry, 10.0 ± 1.9 pc from 2MASS K_s photometry, 9.8 ± 1.9 pc from WISE W1 photometry, 9.8 ± 1.7 pc from WISE W2 photometry, and 9.4 ± 1.9 pc from WISE W3 photometry using the relations from Dupuy & Liu (2012). The uncertainty in the distance estimates comes from the uncertainty in the photometry and the rms from the spectral-type–absolute-magnitude relationships. The mean of these estimates provides a distance of 9.7 ± 1.7 pc, assuming no binarity. We note that if we exclude the J band photometry in the distance estimate due to the photometric confusion flag, the distance increases by 0.02 pc to 9.8 pc. The proximity of W1405+8350 brings the number of very-late L dwarfs within 10 pc to seven, it is a member of the few but growing population of nearby L dwarfs at the L/T transition; see Table 2. Trigonometric parallax measurements are needed for a more reliable distance estimate.

Table 2. Very Late L Dwarfs (⩾L7) within 10 pc

L Dwarf	Spectral Type	Distancea	W1 − W2	W2 − W3	References
	(Optical/NIR)	(pc)
WISE J104915.57−531906.1A	L8 ± 1/L7.5	2.0 ± 0.15b	0.56 ± 0.03c	1.13 ± 0.03c	1, [1, 2], 1
DENIS-P J0255−4700	L8/L9	4.97 ± 0.10d	0.55 ± 0.03	1.01 ± 0.03	3, [4, 5], 6
WISEP J060738.65+242953.4	L8/L9	7.8$^{+1.4}_{-1.2}$e	0.59 ± 0.03	0.81 ± 0.08	7, [8, 8], 7
WISEPA J164715.59+563208.2	⋅⋅⋅/L9 pec (red)	8.6$^{+2.9}_{-1.7}$f	0.52 ± 0.03	1.03 ± 0.10	9, [⋅⋅⋅, 9], 9
WISEP J180026.60+013453.1	⋅⋅⋅/L7.5	8.8 ± 1.0	0.47 ± 0.06	1.21 ± 0.07	10, [⋅⋅⋅, 10], 10
2MASS J02572581−3105523	L8/⋅⋅⋅	9.7 ± 1.3	0.43 ± 0.03	0.99 ± 0.06	4, [4, ⋅⋅⋅], 11
WISE J140533.32+835030.5	L8/L9	9.7 ± 1.7	0.58 ± 0.04	1.00 ± 0.05	8, [8, 8], 8

Notes. WISE colors are from the all-sky data release. ^aBased on spectrophotometric distance estimate unless otherwise noted. ^bBased on least-squares fitting of proper and parallactic motion for the astrometry of WISE J104915.57−531906.1AB. ^cWISE J104915.57−531906.1AB is unresolved in WISE, colors are of the unresolved binary system, the L_o8 ± 1/L_n7.5 (Luhman 2013; Burgasser et al. 2013) primary and the T_o1.5 ± 2/T_n0.5 (Kniazev et al. 2013; Burgasser et al. 2013) secondary. ^dBased on trigonometric parallax. ^eA preliminary parallax of 139 ± 2 mas ($7.19^{+0.11}_{-0.10}$ pc) was determined based on two seasons of data, consistent with the distance estimate from Castro & Gizis (2012). ^fBased on preliminary astrometric measurements from Kirkpatrick et al. (2011), has a spectrophotometric distance estimate of 20.2 pc. References. Discovery, optical and near-infrared spectral type, and distance references: (1) Luhman 2013; (2) Burgasser et al. 2013; (3) Martín et al. 1999; (4) Kirkpatrick et al. 2008; (5) Burgasser et al. 2006a; (6) Costa et al. 2006; (7) Castro & Gizis 2012; (8) This work; (9) Kirkpatrick et al. 2011; (10) Gizis et al. 2011a; (11) Looper et al. 2008b.

Download table as:  ASCIITypeset image

Based on the apparent motion and the estimated distance, W1405+8350 has a tangential velocity of 39 ± 6 km s⁻¹, within range of transverse motions for other L8 dwarfs from Faherty et al. (2009), who quote a median value of 25 km s⁻¹ and a dispersion of 19 km s⁻¹. This v_tan is consistent with that expected for a member of the Galactic thin disk (Faherty et al. 2009). Combined with the radial velocity from the moderate-resolution J band spectrum of −27 ± 14 km s⁻¹, we find a total velocity of 48 ± 15 km s⁻¹, firmly placing it as a member of the galactic thin disk population. Spectral-type–effective-temperature (Looper et al. 2008a) and spectral-type–absolute-bolometric-magnitude (Burgasser 2007a) relationships give a T_eff = 1460 ± 90 K and a log L/L_☉ = −4.56 ± 0.09, where the uncertainty in T_eff comes from the rms in the spectral-type–effective-temperature relation and the uncertainty in log L/L_☉ is from the rms in the spectral-type–absolute-bolometric-magnitude relation. Based on these physical properties, theoretical isochrones from Baraffe et al. (2003) give a range of 0.5 Gyr and 0.03 M_☉ to 10 Gyr and 0.072 M_☉, placing W1405+8350 in the substellar regime, as are all of the latest L dwarfs (Kirkpatrick 2005). A summary of characteristics for W1405+8350 is found in Table 3.

Table 3. Parameters of WISE J140533.32+835030.5

Parameters	WISE J140533.32+835030.5
USNO R.A. (J2000)	14:05:32.57
USNO decl. (J2000)	+83:50:31.7
USNO epoch	2012.23
WISE R.A. (J2000)	14:05:33.32
WISE decl. (J2000)	+83:50:30.6
WISE epoch	2010.25
2MASS R.A. (J2000)	14:05:37.29
2MASS decl. (J2000)	+83:50:24.9
2MASS epoch	2000.13
i (mag)a	20.56 ± 0.16
USNO IC (mag)	18.93 ± 0.06
USNO z (mag)	17.50 ± 0.04
2MASS J (mag)	14.63 ± 0.04
2MASS H (mag)	13.49 ± 0.04
2MASS Ks (mag)	12.85 ± 0.03
WISE W1 (mag)	12.08 ± 0.03
WISE W2 (mag)	11.50 ± 0.03
WISE W3 (mag)	10.50 ± 0.04
WISE W4 (mag)	>9.68
IC − z (mag)	1.43 ± 0.07
IC − J (mag)	4.30 ± 0.07
i − z (mag)	3.06 ± 0.16
i − J (mag)	5.93 ± 0.16
z − J (mag)	2.87 ± 0.06
J − Ks (mag)	1.78 ± 0.06
W1 − W2 (mag)	0.58 ± 0.04
H − W2 (mag)	1.99 ± 0.05
Ks − W2 (mag)	1.35 ± 0.04
μαcos(δ) (mas yr−1)	−630 ± 10
μδ (mas yr−1)	570 ± 20
Total motion ('' yr−1)	0.85 ± 0.02
Spectral type (Opt/NIR)	L8/L9
Distance (pc)	9.7 ± 1.7
vtan (km s−1)	39 ± 7
vrad (km s−1)	−27 ± 14
vtot (km s−1)	48 ± 16
Teff (K)	1460 ± 90
log L/L☉	−4.56 ± 0.09
Mass at 0.5 Gyr (M☉)	0.030
Mass at 10 Gyr (M☉)	0.072

Notes. WISE data is from the all-sky data release. ^aTransformed from I_C using transformation derived in Section 2.4.

Download table as:  ASCIITypeset image

Approximately 20% of L and T dwarfs are resolved as very-low-mass binaries, with the resolved binary fraction of L/T transition dwarfs (L7-T3.5) double that of other L and T dwarfs (Burgasser et al. 2006b). Field binaries are primarily equal brightness/mass systems in tightly bound orbits (<20 AU), where the separation of binary systems peaks at <10 AU (Allen 2007; Burgasser et al. 2007b). A secondary to W1405+8350 of equal or earlier spectral type (≲L8) would have been detected at ≳15 AU based on the FWHM (≈15) of the USNO z band image. If W1405+8350 was an unresolved binary system, for example, consisting of two L9 dwarfs, it would push the spectrophotometric distance estimate to 13.8 pc. The highest resolution imaging/spectroscopy is warranted to search for a companion to W1405+8350. The sensitivity of current imaging surveys begins to fall off at separations of ≲ 3–4 AU, where there is a model predicted frequency peak of binarity (Allen 2007). Close L dwarfs such as W1405+8350, if found to have companions, will play a crucial role in extending binaries into this regime.

4.2.1. Near-infrared and Optical Spectroscopy

Figure 7 shows the SpeX NIR spectrum of WISE J040137.21+284951.7 (W0401+2849) (black) compared to NIR standards (red) and a reference dwarf (blue) also observed with SpeX. We use the goodness-of-fit χ² as defined in Burgasser (2007b) to compare W0401+2849 to standard/reference spectra. W0401+2849 fits equally well the NIR standards Kelu-1 (L_n2) (χ² = 0.25) and 2MASSW J1506544+132106 (L_n3) (χ² = 0.25), with an excellent fit (χ² ⩽ 0.10) to the reference dwarf DENIS-P J1058.7−1548 (L_o3/L_n3) (χ² = 0.08). We use the spectral indices defined in Burgasser et al. (2006a) and the spectral-index relation from Burgasser (2007a, their Table 3) to obtain spectral indices for W0401+2849 of L3.1±0.8 (H₂O–J), L2.6 ± 1.0 (H₂O–H), and L3.6 ± 1.1 (CH₄–K), for a mean spectral type of L3.1 ± 1.0; see Table 4 for details of the NIR spectral indices of W0401+2849. The average disagreement between the spectral-index relation and published classifications is σ = 1.1 subtypes for L dwarfs, systematically later for early L dwarfs (Burgasser 2007a; Burgasser et al. 2010). The spectral-index relation gives a mean spectral type of L2.7 ± 1.0 and L3.4 ± 1.0 for the NIR standards Kelu-1 (L_n2) and 2MASSW J1506544+132106 (L_n3), respectively, too late on average by about half a spectral subtype. We adopt a NIR spectral type of L2.5 ± 0.5 for W0401+2849.

Zoom In Zoom Out Reset image size

Table 4. L Dwarf Near-infrared Spectral Indices

Object	H2O–J	H2O–H	CH4–K	Near-infrared Spectral Type
				Indexa	Direct
WISE J040137.21+284951.7	0.838 (L3.1)	0.800 (L2.6)	1.020 (L3.6)	L3.1 ± 1.0	L2.5 ± 0.5
WISE J040418.01+412735.6	0.834 (L3.2)	0.801 (L2.5)	1.064 (L1.8)	L2.5 ± 1.0	L2 pec (red)
WISEP J060738.65+242953.4	0.695 (L7.5)	0.693 (L6.8)	0.835 (L8.9)	L7.7 ± 1.0	L9
WISE J062442.37+662625.6	0.904 (L1.3)	0.836 (L1.1)	1.059 (L2.0)	L1.5 ± 1.0	L1

Note. ^aSpectral indices defined by Burgasser et al. (2006a), with corresponding spectral type determined from the spectral-index relation in Burgasser (2007a). Final index spectral type is the mean of the individual spectral indices.

Download table as:  ASCIITypeset image

Figure 6 shows the optical spectrum for W0401+2849 without telluric corrections. The best fit to W0401+2849 is the L3 optical standard 2MASSW J1146345+223053, we adopt an optical spectral type of L3 for W0401+2849.

4.2.2. Distance and Physical Properties

4.3.1. Near-infrared and Optical Spectroscopy

Figure 8 shows the SpeX NIR spectrum of WISE J040418.01+412735.6 (W0404+4127) (black) compared to NIR standards (red) and a reference dwarf (blue). The spectrum of W0404+4127 matches the J band of the L2, L3, and L4 NIR standards equally well, while the FeH absorption in the H band identifies it as an early L dwarf (L2-L4). With excess flux in the H and K bands the spectrum is clearly red, with the slope of the spectrum similar to the L5 NIR standard. The goodness-of-fit χ² between W0404+4127 and the (L1-L4) NIR standards was determined in an alternative manner. The spectra were normalized to the J band, H band, and K band, and the χ² test was applied to those respective regions, with the χ² shown being the sum of the three individual χ² values. The L2, L3, and L4 NIR standards fit the J band approximately equally well (χ² = 0.10, 0.09, 0.10, respectively), the L2 fitting the H band best, the L3 fitting the K band best, with the sum of χ² for the J, H, and K giving the NIR standard Kelu-1 (L_n2) (χ² = 0.22) the best fit to W0404+4127. The spectral-index relation gives L3.2 ± 0.8 (H₂O–J), L2.5 ± 1.0 (H₂O–H), and L1.8 ± 1.1 (CH₄–K), for a mean spectral type of L2.5 ± 1.0; see Table 4 for details of the NIR spectral indices of W0404+4127. W0404+4127 has a J − K_s = 1.73 ± .04, ∼0.2 above the average for L2 dwarfs (Faherty et al. 2009), J − K_s = 1.52 ± 0.20 (where the uncertainty is the 1σ standard deviation), typical of peculiar red L dwarf discoveries from Kirkpatrick et al. (2010). The best fit template to W0404+4127 is 2MASS J08234818+2428577, an L3 optical, with no NIR spectral classification found in the literature, it appears to be an overlooked red peculiar L dwarf.

Zoom In Zoom Out Reset image size

Figure 6 shows the optical spectrum for W0404+4127 without telluric corrections. The best fit to W0404+4127 is the L2 optical standard Kelu-1, we adopt an optical spectral type of L2 for W0404+4127.

4.3.2. Unusually Red L Dwarf?

4.3.3. Unresolved Binary?

We investigate whether the red spectrum of W0404+4127 could be reproduced by the combined light spectrum of two unresolved L dwarfs. We use the technique of binary spectral template matching following the procedures from Burgasser (2007b), using L dwarf NIR standards from Kirkpatrick et al. (2010). The primary and the secondary are scaled based on their flux contribution according to the M_K-spectral-type relation of Burgasser (2007a), with the source and light combined spectra being normalized to the J band. Figure 9 shows the best fit composite spectrum for W0404+4127, an L2 primary and an L7 secondary, with a fit of χ² = 0.40 determined in the standard manner. The composite fit to the H band (χ² = 0.09, normalized to the J band) is almost as good as the best fitting single Kelu-1 (L_n2) (χ² = 0.06, normalized to the H band), mostly retaining the FeH absorption feature. However, the fit to the K band is not red enough. Although the composite provides an improvement over the best fit single L dwarf, it does not adequately reproduce the red spectrum of W0404+4127. If the flux contribution from the L7 were marginally more luminous relative to the L2, qualitatively this would increase the slope of the composite, potentially providing a better fit. We examine the L2+L7 composite and the effect of including the uncertainty in the M_K-spectral-type relation of Burgasser (2007a), with an uncertainty in the relation of σ = 0.26 mag. Allowing the uncertainty to cause M_K to be brighter for the L7 at intervals of 0.5σ provides an improved fit, where the best fit occurs at 2.5σ with χ² = 0.23. The best fit no longer has a deficit in K band flux (except for a slight deficit at longer wavelengths), however, it does have a peakier H band (χ² = 0.09) and an unrealistic 82% increase in flux of the K band of the L7. The L7 2MASSI J0103320+193536 has a J–K_s = 2.14 ± 0.10, ∼0.33 above the average for L7 dwarfs (Faherty et al. 2009). In order to explain the red spectrum of W0404+4127 as an unresolved binary, the L7 secondary may need to be a red peculiar L dwarf itself.

Zoom In Zoom Out Reset image size

Kirkpatrick et al. (2010) found on average large space velocities from a small sample of red L dwarfs, indicating a kinematically older population, in contrast with low gravity (youth) or high metallicity. The kinematics of W0404+4127 are consistent with the higher transverse velocities of red L dwarfs from Kirkpatrick et al. (2010), suggesting an older age, however, radial velocity measurements are needed to confirm a large space velocity. We adopt a NIR spectral type of L2 pec (red) and add W0404+4127 to the class of unusually red L dwarfs.

4.3.4. Distance and Physical Properties

4.4.1. Near-infrared and Optical Spectroscopy

Figure 10 shows the SpeX NIR spectrum of WISE J062442.37+662625.6 (W0624+6626) (black) compared to NIR standards (red) and a reference dwarf (blue). W0624+6626 best fits the L1 NIR standard 2MASSW J2130446−084520, with a very slight lack of flux in the H and K band compared to 2MASSW J2130446−084520. The best fitting template to W0624+6626 is DENIS-P J170548.38−051645.7 (χ² = 0.10), classified as an L_n4 by Kendall et al. (2004), in contradiction to the spectrum shown here, and given a NIR SpeX spectral type of L1.0 by Burgasser et al. (2010). The spectral-index relation gives L1.3 ± 0.8 (H₂O–J), L1.1 ± 1.0 (H₂O–H), and L2.0 ± 1.1 (CH₄–K), for a mean spectral type of L1.5 ± 1.0, within uncertainty of the direct fit; see Table 4 for details of the NIR spectral indices of W0624+6626. The spectra-index relation gives a mean spectral type of L1.0 ± 1.0, L1.4 ± 1.0, and L2.7 ± 1.0 for the NIR standards 2MASS J03454316+2540233 (L_n0), 2MASSW J2130446−084520 (L_n1), and Kelu-1 (L_n2), respectively, too late on average by about half a spectral subtype. We adopt a NIR spectral type of L1 for W0624+6626.

Zoom In Zoom Out Reset image size

Figure 6 shows the optical spectrum for W0624+6626 without telluric corrections. The best fit to W0624+6626 is the L1 optical standard 2MASSW J1439284+192915, we adopt an optical spectral type of L1 for W0624+6626.

4.4.2. Distance and Physical Properties

Figure 11 shows the SpeX NIR spectrum of W0607+2429 (black) compared to L/T transition dwarf NIR standards (red) and a reference dwarf (blue). W0607+2429 is a very good fit to the L9 NIR standard DENIS-P J0255−4700. The spectral-index relation gives L7.5 ± 0.8 (H₂O–J), L6.8 ± 1.0 (H₂O–H), and L8.9 ± 1.1 (CH₄–K), for a mean spectral type of L7.7 ± 1.0. We note that the weak absorption at the 1.15 μm H₂O bands gives a slightly earlier H₂O–J spectral-index estimate. The index spectral type is slightly earlier than the direct spectral type of L9, but is consistent within 2σ of the uncertainty. The spectral-index relation produces early spectral types compared to published classifications for late L dwarfs (Burgasser 2007a; Burgasser et al. 2010). The spectral-index relation gives a mean spectral type of L7.1 ± 1.0 for 2MASSW J1632291+190441 (L_o8/L_n8), L8.0 ± 1.0 for DENIS-P J0255−4700 (L_o8/L_n9), L8.7 ± 1.0 for 2MASSI J0328426+230205 (L_o8/L_n9.5), and L9.5 ± 0.8 for SDSS J120747.17+024424.8 (L_o8/T_n0), all within the uncertainty of their published NIR spectral types but systematically earlier by a mean subtype of 0.8. Table 4 provides details of the NIR spectral indices of W0607+2429.

Zoom In Zoom Out Reset image size

The spectrum of W0607+2429 shows weaker absorption at the 1.15 μm H₂O bands compared to DENIS-P J0255−4700, but is similar to 2MASSW J1632291+190441 and 2MASSI J0328426+230205; it may be that DENIS-P J0255−4700 has unusually strong 1.15 μm H₂O bands and W0607+2429 is normal. W0607+2429, similar to DENIS-P J0255−4700 and 2MASSI J0328426+230205, show weak signs of the 2.2 μm CH₄ band that is typical of the latest-type L dwarfs (Burgasser 2007b). Interestingly, W0607+2429 has "ambiguous" CH₄ absorption at 1.6 μm in the H band compared to DENIS-P J0255−4700, with a downward sloping peak compared to the more rounded peak of DENIS-P J0255−4700. The CH₄ absorption in the H band is rather similar to 2MASSI J0328426+230205 (L_o8/L_n9.5), but not as deep as SDSS J120747.17+024424.8 (L_o8/T_n0). We consider the unambiguous detection of CH₄ in the H band to be that of the T0 standard (Burgasser et al. 2006a) SDSS J120747.17+024424.8, where the unambiguous detection at low resolution of CH₄ at the H and K bands is the defining characteristic of the T dwarf spectral class (Kirkpatrick 2005). The methane index in the H band is CH₄–H = 1.00 for W0607+2429, 1.08, 1.03, 1.01, and 0.95 for 2MASSW J1632291+190441 (L_n8), DENIS-P J0255−4700 (L_n9), 2MASSI J0328426+230205 (L_n9.5), and SDSS J120747.17+024424.8 (T_n0), respectively; with CH₄–H >0.97 indicating that CH₄ absorption is not "present" in the H band (Burgasser 2007a). We note that 2MASSI J0328426+230205 is suspected of being an unresolved binary (Burgasser et al. 2010), likely due to the same "ambiguous" CH₄ absorption found in the H band of W0607+2429.

Figure 6 shows the optical spectrum for W0607+2429 without telluric corrections. The best fit to W0607+2429 is the L8 optical standard 2MASSW J1632291+190441, we adopt an optical spectral type of L8 for W0607+2429.

Approximately 20% of L and T dwarfs are resolved as very-low-mass binaries, with the resolved binary fraction of L/T transition dwarfs (L7-T3.5) double that of other L and T dwarfs (Burgasser et al. 2006b). Numerous L dwarfs with spectra exhibiting features of methane in the H band have been shown to be unresolved L/T binaries based on the convention of empirical binary templates (Burgasser et al. 2010; Burgasser 2007b). The "ambiguous" CH₄ absorption in the H band is suggestive that W0607+2429 could be an unresolved L/T binary. The combined light spectrum from a late L dwarf and an early T dwarf can give a direct/index spectral type of L9 (Looper et al. 2007; Burgasser 2007a). Based on synthetic spectral binary templates from Looper et al. (2007), an unresolved binary of an L8+T0 (NIR) would give a direct comparison spectral type of L9, and an L8+T1, L9+T0, and L9+T1 would yield an L9.5 spectral type.

We investigate whether the CH₄ absorption in the H band of W0607+2429 could be reproduced by the combined light spectrum of a primary L dwarf and a secondary T dwarf. We construct composite spectra in the same manner as for W0404+4127, however, here we use L and T NIR standards from Kirkpatrick et al. (2010); Burgasser et al. (2006a); with the exception of the T1 NIR standard, where we use SDSS J015141.69+124429.6 from Burgasser (2007a). Figure 12 shows the best fit composite spectrum, the only binary template with χ² < 0.50, an L8+T0. The combined light spectrum of the L8+T0 (χ² = 0.34) reproduces the CH₄ absorption feature in the H band, albeit slightly deeper, but does not provide a better fit than the single L9 standard (χ² = 0.22), with the flux in the H and K band too weak. Combinations of an L8 primary and later secondary T dwarfs result in CH₄ absorption that is too deep and flux in the H and K band increasingly weaker. Likewise, combinations of an L7 primary and T dwarf secondaries do not reproduce the CH₄ in the H band and have excess flux in the K band. For the L8+T0 composite, if the flux contribution from the T0 were lowered relative to the L8, qualitatively this would increase the H and K band flux, potentially providing a better fit. We examine the L8+T0 composite and the effect of including the uncertainty in the M_K-spectral-type relation of Burgasser (2007a). Allowing the uncertainty to cause M_K to be fainter for the T0 from 0σ to 1σ at intervals of 0.25σ results in an increase in the H and K band flux and an improving fit from χ² = 0.34 to χ² = 0.19. However, in the process the J band of the T0 is reduced relative to the L8 such that at 1σ the J band peak of the L8 and the T0 are equivalent, the characteristic J band bump between late L and early T dwarfs has vanished, and even slightly reversed. If an uncertainty of 0.75σ is allowed in the M_K-spectral-type relation causing the T0 to be fainter relative to the L8, then the L8+T0 composite provides a fit of χ² = 0.21, slightly better than the single L9 NIR standard (χ² = 0.22). This improved fit comes with a J band bump that is almost absent and a 16% reduction in flux of the K band of the T0. The best fit occurs at 1.25σ and 1.5σ, with χ² = 0.18, where the J band bump has reversed. The binary templates do not lend credence to W0607+2429 being an unresolved L/T transition binary, nor do they exclude it. It may be that the ambiguous appearance of CH₄ absorption is a natural phenomenon, in this case the onset of CH₄ in the H band as early as an L9 rather than an L9.5. The highest resolution imaging and radial velocity measurements are warranted to further investigate the potential binarity of W0607+2429.

Zoom In Zoom Out Reset image size

The L9 NIR standard DENIS-P J0255−4700 shows weak CH₄ absorption in the H band in the medium resolution SpeX spectrum (Cushing et al. 2005), but no "distinct" CH₄ absorption in the low resolution SpeX spectrum (Burgasser et al. 2006a). Cushing et al. (2005) argued the case for whether DENIS-P J0255−4700 should be classified as a T dwarf based on the definition of Geballe et al. (2002), who define the boundary between the L/T transition as the earliest appearance of methane absorption in both the H and K bands. Burgasser et al. (2006a) classified DENIS-P J0255−4700 as an L9 based on an analogous definition by Burgasser et al. (2006a), that the T dwarf sequence begins with the detection of CH₄ absorption in the H band at "low" resolution. Dwarfs classified as late L that show "distinct" CH₄ absorption at low resolution akin to W0607+2429 in the H band, such as 2MASSI J0328426+230205, are generally classified as L_n9.5. W0607+2429 does not fit the SED of L9.5 dwarfs well with excess flux in the H and K bands comparatively; compared to all templates in the SpeX library the best fit to W0607+2429 being the L9 NIR standard DENIS-P J0255−4700. By an analogous argument to DENIS-P J0255−4700, one could argue for the classification of W0607+2429 as a T dwarf based on the presence of "distinct" CH₄ absorption in the H band at low resolution, meeting the definition of Burgasser et al. (2006a) for a T dwarf. However, the overall SED of W0607+2429 is clearly earlier than T0, in excellent agreement with DENIS-P J0255−4700. W0607+2429 shows the earliest onset of CH₄ in the H band compared to all of the L9 dwarfs in the SpeX library that have an overall SED consistent with the L9 NIR standard, DENIS-P J0255−4700; with several objects classified as L9 showing CH₄ in the H band whose H and K bands are weaker than the L9 NIR standard and not consistent with the overall SED. W0607+2429 may be the catalyst for a modification to what is defined as a T dwarf if it is indeed a single object, with the onset of "distinct" CH₄ in the H band as early as L9; and may help to shape our understanding of the L/T transition.

Based on the direct fit and the systematically early index spectral classification, we adopt an NIR spectral type of L9 for W0607+2429. The adopted optical and NIR spectral type confirm the optical spectral type estimate from Castro & Gizis (2012), illustrating the utility and accuracy of color–color diagrams in estimating the spectral type of normal colored very late L dwarfs. Discrepant optical/NIR spectral types and CH₄ absorption features in the H band are characteristics of unresolved L/T binary systems (Cruz et al. 2004; Burgasser et al. 2005, 2010; Burgasser 2007b). The complementary optical and NIR spectral types of W0607+2429 provide support that it is a single object.