Redder than Red: Discovery of an Exceptionally Red L/T Transition Dwarf

In an effort to refine the astrometry and photometry of CWISE J0506+0738, we observed it with the J_MKO filter on the infrared Wide-Field Camera (WFCAM; Casali et al. 2007) on UKIRT on 2022 September 20. Observations were performed using a 3 × 3 microstepping pattern, with the resulting nine images interleaved (Dye et al. 2006) to provide improved sampling over that of a single WFCAM exposure. The microstepping sequence was repeated 5 times, resulting in 45 single exposures, each lasting 20 s, for a total exposure time of 900 s. We reregistered the world coordinate system of each interleaved frame using the Gaia DR3 catalog (Gaia Collaboration et al. 2022). Images were then combined using the imstack routine from the CASUTOOLS package ¹⁴ (Irwin et al.2004). The position and photometry of CWISE J0506+0738 were extracted using the CASUTOOLS imcore routine.

Combining the position of this J-band observation with the UHS K-band observation, we calculated proper motion components of μ_α = 31.5 ± 2.6 mas yr⁻¹ and μ_δ = −82.7 ± 2.7 mas yr⁻¹. CatWISE 2020 reports proper motions of μ_α = 44.2 ± 7.9 mas yr⁻¹ and μ_δ = −97.5 ± 8.4 mas yr⁻¹ (with offset corrections applied according to Marocco et al. 2021). The proper motion calculated from our UKIRT observations is significantly more precise than the proper motion measurements from CatWISE 2020, and we adopt the former for our analysis.

We measure a J_MKO − band magnitude of 18.487 ± 0.017 mag from these observations, which is >2σ brighter than the value from the UHS (18.76 ± 0.10 mag). To verify our measured photometry, we compared the photometry for other sources found to have similar magnitudes (18.4 <J < 18.6 mag) in our images to UHS values. We found a median J-band difference for the 52 objects in this sample to be −0.03 mag, with a median absolute deviation of 0.07 mag, showing that differences as large as those measured for this object (0.27 mag) are relatively rare. The origin of the difference between these J-band measurements is unclear though we note that variability may be a contributing factor as young (and red) objects are often found to have larger amplitude variability than field-age objects with similar spectral types (e.g., Vos et al. 2022a). While this new J-band measurement results in bluer (J − K)_MKO = 2.97 ± 0.03 mag and J_MKO − W2 = 4.94 ± 0.02 mag colors, they both remain significantly redder than those of any previously identified free-floating brown dwarf.

All UKIRT photometry and astrometry for CWISE J0506+0738 are provided in Table 1.

Table 1. Properties of CWISE J050626.96+073842.4

Parameter	Value	References
R.A. (°) (epoch = 2022.7) a	76.6124377	1
Decl. (°) (epoch = 2022.7) a	7.6449299	1
R.A. (°) (epoch = 2017.8) a	76.6123885	2
Decl. (°) (epoch = 2017.8) a	7.6450716	2
μα (mas yr−1)	31.5 ± 2.6	1
μδ (mas yr−1)	−82.7 ± 2.7	1
d b (pc)	${32}_{-3}^{+4}$	1
RV (km s−1)	+16.3${}_{-7.7}^{+8.8}$	1
JMKO (mag)	18.487 ± 0.017	1
KMKO (mag)	15.513 ± 0.022	2
W1 (mag)	14.320 ± 0.015	3
W2 (mag)	13.552 ± 0.013	3
Sp. Type	L8γ–T0γ	1

Notes.

^a R.A. and decl. values are given in the International Celestial Reference System. ^b Photometric distance estimate based on the UHS K_MKO-band magnitude and the absolute magnitude-spectral type relation in Dupuy & Liu (2012) (see Section 5.3).

References. (1) This work; (2) UHS DR2 (Dye et al. 2018; J. Bruursema et al. 2023, in preparation); (3) CatWISE 2020 (Marocco et al. 2021).

Download table as:  ASCIITypeset image

CWISE J0506+0738 was observed with the Near-Infrared Echellette Spectrometer (NIRES; Wilson et al. 2004) mounted on the Keck II telescope on UT 2022 January 19. NIRES provides a resolution λ/Δλ ≈ 2700 over five cross-dispersed orders spanning a wavelength range of 0.9–2.45 μm. CWISE J0506+0738 was observed in four 250 s exposures nodded in an ABBA pattern along the slit, which was aligned with the parallactic angle, for a total on-source integration time of 1000 s. The spectrum was extracted using a modified version of the SpeXTool package (Vacca et al. 2003; Cushing et al. 2004), with the A0 V star HD 37887 (V = 7.67) used for telluric correction. The large J − K color of CWISE J0506+0738 resulted in significant signal-to-noise ratio (S/N) differences across the final reduced spectrum, with an S/N ∼ 25 at the J-band peak (∼1.3 μm) and an S/N ∼ 200 at the K-band peak (∼2.2 μm).

The interband flux calibration for Keck/NIRES orders is occasionally skewed by seeing or differential refraction slit losses. In particular, there is a gap between the third (K-band) and fourth (H-band) orders spanning 1.86 to 1.89 μm, ¹⁵ and the overlap between the fourth and fifth (J-band) orders lies in a region of strong telluric and stellar H₂O absorption. We therefore rescaled the resulting spectrum to have a J − K synthetic color consistent with UKIRT J-band and UHS K-band photometry by applying small multiplicative constants to the H- and K-band portions of the spectrum. The final reduced spectrum is shown in Figure 2.

Zoom In Zoom Out Reset image size

As with many of the known, young, late-type red L dwarfs, none of the L dwarf spectral standards (Kirkpatrick et al. 2010; Cruz et al. 2018) provide a suitable match to the near-infrared spectrum of CWISE J0506+0738. The best match to the J-band portion of the spectrum is the L7 standard 2MASSI J0825196+211552 (Kirkpatrick et al. 1999; Cruz et al. 2018), which is shown in the upper panel of Figure 2. CWISE J0506+0738 shows much stronger H₂O absorption around 1.1 μm, a feature commonly seen in low-gravity L dwarfs. This comparison also shows how red CWISE J0506+0738 is compared to a normal, field-age/field-gravity late L dwarf. The lower panel of Figure 2 shows a comparison of CWISE J0506+0738 with PSO J318.5338−22.8603 (Liu et al. 2013a), which is typed as L7 VL-G. These two objects match relatively well across the J-band portion of the spectrum though the extreme redness of CWISE J0506+0738 can still be seen in this comparison via the mismatch in the H- and K-band portions of their spectra.

We also note that the spectrum of CWISE J0506+0738 has a noticeable absorption feature at the H-band peak. There is also a second, less-pronounced absorption feature present in the K-band portion of CWISE J0506+0738's spectrum between 2.2 and 2.3 μm. While we cannot a priori rule out systematic noise or a data reduction artifact for these features, we note that no similar features have been seen in Keck/NIRES spectra of L dwarfs obtained and reduced by our group (e.g., Meisner et al. 2021; Schapera et al. 2022; Softich et al. 2022; Theissen et al. 2022). We also note that these features occur at the approximate locations of CH₄ absorption seen in model spectra of low-surface gravity brown dwarfs with effective temperatures ≲1400 K. Figure 3 compares solar-metallicity model spectra from Marley et al. (2021) with fixed low-surface gravities (log(g) = 3.5) and varying effective temperatures. Prominent methane absorption features can be seen in the H and K bands for T_eff ≲ 1400 K. While these models are informative for (potentially) identifying the source of some of the absorption features seen in the spectrum of CWISE J0506+0738, we were unable to find any models that successfully reproduced the overall shape of CWISE J0506+0738's spectrum, similar to previous studies of young brown dwarfs (e.g., Manjavacas et al. 2014).

Zoom In Zoom Out Reset image size

The presence of CH₄ in the H- and K-band peaks of CWISE J0506+0738's spectrum would suggest that this source is early T dwarf (Burgasser et al. 2006) although these features are fairly weak in strength. Charnay et al. (2018) showed that the presence of clouds can greatly reduce the abundance of CH₄ in the photospheres of low-gravity objects, a possible explanation for the absence of CH₄ bands in the spectra of 2M1207b and HR8799bcd (Barman et al. 2011a, 2011b; Konopacky et al. 2013). If the same effect holds here, it would argue for a particularly low temperature for CWISE J0506+0738, below that of the T_eff ≈ 1200 K planetary-mass L dwarf PSO J318.5338−22.8603 and VHS 1256−1257B, which originally showed no indication of CH₄ absorption in the 1–2.5 μm region. ¹⁶ (Liu et al.2013a; Gauza et al. 2015). These two sources do have detectable absorption in the 3.3 μm ν₃ CH₄ fundamental band (Miles et al. 2018), and cloud scattering opacity is likely responsible for muting the 1.6 and 2.2 μm bands in these red L dwarfs (Charnay et al. 2018; Burningham et al. 2021). Indeed, it has been noted previously that PSO J318.5338−22.8603 is just on the warmer side of the transition to CH₄ becoming the dominant carbon-bearing molecule in its atmosphere (Tremblin et al. 2017). We tentatively assert that both H- and K-band features in the spectrum of CWISE J0506+0738 are due to CH₄ absorption, which may be tested with more detailed analysis (e.g., atmospheric retrievals; Burningham et al. 2017, 2021) and higher S/N moderate-resolution data. Given the similarity of the J-band portion of CWISE J0506+0738's spectrum to PSO J318.5338−22.8603 (L7 VL-G) and the likely detection of CH₄ in the H and K bands, we assign a near-infrared spectral type of L8γ–T0γ to CWISE J0506+0738, where the γ signifies very low surface gravity (Kirkpatrick2005).

The characterization of brown dwarfs and planetary-mass objects as "low surface gravity" or "young" typically arises from gravity-sensitive (or more specifically, photosphere pressure-sensitive) spectral features quantified by spectral indices (e.g., Steele & Jameson 1995; Martin et al. 1996; Luhman et al. 1997; Gorlova et al. 2003; McGovern et al. 2004; Kirkpatrick et al. 2006; Allers et al. 2007; Manjavacas et al. 2020). Many of these spectral indices, however, are designed for optical spectra (e.g., Cruz et al. 2009) or are only applicable to objects with spectral types earlier than ∼L5 (e.g., Allers & Liu 2013; Lodieu et al. 2018). The H-cont index is a gravity-sensitive index defined in Allers & Liu (2013) that is one of the few gravity-sensitive indices applicable to spectral types later than L5. This index is designed to approximate the slope of the blue side of the H-band peak, with low-gravity objects exhibiting a much steeper slope than field-age brown dwarfs. However, this index is defined using a band centered at 1.67 μm, which is where a feature potentially attributable to CH₄ occurs in our spectrum. Thus the H-cont index does not provide an accurate assessment of the slope of the blue side of the H-band peak for this object.

We have created a modified slope index for the blue side of the H-band peak by computing a simple linear least-squares fit to the 1.45–1.64 μm region after normalizing to the J-band peak between 1.27 and 1.29 μm. We measured this slope (normalized flux/μm) for several late-L and early-T dwarfs, both field and young association members, as shown in Figure 4. We note that the largest slope for the entire sample belongs to WISE J173859.27+614242.1, an object that has been difficult to classify (Mace et al. 2013), but is most consistent with an extremely red L9 (Thompson et al. 2013). It is unclear if this object is young, has an extremely dusty photosphere, or both. For typical L7-T0 dwarfs, H-slope values for field objects range from 2 to 4, while equivalently classified young L dwarfs have values that range over 3–5. For CWISE J0506+0738, we find a slope of 4.38, significantly larger than field-age late-L dwarfs. The known population of young, very red L dwarfs similarly has larger H-slope values than their field-age counterparts.

Zoom In Zoom Out Reset image size

Schneider et al. (2014) also showed that the H₂(K) index defined in Canty et al. (2013) could distinguish young, low-gravity late-L dwarfs from the field late-L population. The H₂(K) index determines the slope of the K band between 2.17 and 2.24 μm. CWISE J0506+0738 has an H₂(K) value of 1.030, which is again consistent with the known population of low-gravity late-type L dwarfs (1.029 ≤ H₂(K) ≤ 1.045) compared to field-age L6–L8 brown dwarfs (H₂(K) ≳ 1.05).

Another spectral feature that has been used to distinguish low-surface gravity late-L dwarfs are the K I absorption lines between 1.1 and 1.3 μm (McGovern et al. 2004; Allers & Liu 2013; Miles et al. 2022). Our Keck/NIRES spectrum does not have sufficient S/N around the J-band peak to investigate these lines. A higher S/N spectrum would help to ensure no ambiguity regarding the surface gravity of CWISE J0506+0738.

The resolution of the Keck/NIRES data is sufficient to obtain a coarse measure of the radial velocity (RV) of CWISE J0506+0738, particularly in the vicinity of strong molecular features. We followed a procedure similar to that described in (Burgasser et al. 2015; see also Blake et al. 2010; Hsu et al. 2021), forward-modeling the wavelength-calibrated spectrum prior to telluric correction in the 2.26–2.38 μm region. This spectral band contains the prominent 2.3 μm CO 2–0 band present in L dwarf spectra, as well as strong telluric features that allow refinement of the spectral wavelength calibration (see Newton et al. 2014). We used a T_eff = 1300 K, $\mathrm{log}g$ = 4.5 dex (cgs) BTSettl atmosphere model (M[λ]) from Allard et al. (2012), which provides the best match to the CO band strength, and a telluric absorption model (T[λ]) from Livingston (1991). We forward-modeled the data (D[λ]) using four parameters: the barycentric RV of the star (RV_⊕), the strength of telluric absorption (α), the instrumental Gaussian broadening profile width (σ_broad ), and the wavelength offset from the nominal SpeXtool solution (Δλ):

$$\begin{eqnarray}&&D[\lambda ]=\left(M[{\lambda }^{* }+{\rm{\Delta }}\lambda ]\times T{\left[\lambda +{\rm{\Delta }}\lambda \right]}^{\alpha }\right)\otimes {\kappa }_{G}({\sigma }_{\mathrm{broad}}),\end{eqnarray} \tag{ 1 }$$

with λ* = λ(1 + RV_⊕/c) accounting for the radial motion of the star and κ_G representing the Gaussian broadening kernel. Preliminary fits that additionally included rotational broadening of the stellar spectrum indicated that this parameter was equal to the instrumental broadening and is likely unresolved ($v\sin i$ ≲ 65 km s⁻¹), so it was ignored in our final fit.

After an initial "by-eye" optimization of parameters, we used a simple Markov Chain Monte Carlo (MCMC) algorithm to explore the parameter space, evaluating goodness of fit between model and data using a χ² statistic. Figure 5 displays the posterior distribution of our fit parameters after removing the first half of the MCMC chain ("burn-in"), which is normally distributed. There is a small correlation between RV_⊕ and Δλ, which is expected given that stellar and telluric features are intermixed in this region. This correlation increases the uncertainties of these parameters. We find that the best-fit model from this analysis is an excellent match to the NIRES spectrum, with residuals consistent with uncertainties. After correction for barycentric motion (−19.2 km s⁻¹), we determine a heliocentric RV of +16.3${}_{-7.7}^{+8.8}$ km s⁻¹ for CWISE J0506+0738.

Zoom In Zoom Out Reset image size

CWISE J0506+0738 has exceptionally red colors compared to the known brown dwarf population. Figure 6 highlights this by comparing CWISE J0506+0738 to other UHS DR2 L and T dwarfs (A. Schneider et al. 2023, in preparation) and red L dwarfs not covered by the UHS survey. Table 2 summarizes photometric and spectral type information for all known free-floating L dwarfs with J − K colors greater than 2.2 mag. All photometry is on the Mauna Kea Observatories (MKO) system and comes from Liu et al. (2016), Best et al. (2021), or the VISTA Hemisphere Survey (VHS; McMahon et al. 2013). WISE J173859.27+614242.1 has no near-infrared MKO photometry in the literature or in available catalogs. For this source, we used its low-resolution near-infrared spectrum published in Mace et al. (2013) normalized to its most precise K-band photometric measurement (2MASS K_S; Skrutskie et al. 2006), and then computed synthetic J_MKO and K_MKO photometry. Even among known red L dwarfs, CWISE J0506+0738 stands out as exceptionally red, being ∼0.3 mag redder in both (J − K)_MKO and J_MKO − W2 color than all other known free-floating L dwarfs.

Zoom In Zoom Out Reset image size

Directly imaged planetary-mass companions also have exceptionally red near-infrared colors. Some of the L-type companions (Table 2) do not have WISE W1 (3.4 μm) and W2 (4.6 μm) photometry but have equivalent Spitzer/IRAC photometry in ch1 (3.6 μm) and ch2 (4.5 μm). For HD 203030B, we use J- and K-band photometry from Metchev & Hillenbrand (2006) and Miles-Páez et al. (2017) and convert Spitzer/IRAC ch1 and ch2 photometry from Martinez & Kraus (2022) using the Spitzer-WISE relations from Kirkpatrick et al. (2021). For VHS 1256−1257B, we use J- and K-band photometry from Gauza et al. (2015) and convert Spitzer/IRAC ch2 photometry from Zhou et al. (2020) to W2 using the Kirkpatrick et al. (2021) relation. We chose not to use the published W1 photometry of VHS 1256−1257B from Gauza et al. (2015) because of its large uncertainty (0.5 mag). For BD+60 1417B, all photometry comes directly from Faherty et al. (2021). Both HD 203030B and BD+60 1417B are included in both panels of Figure 6, while VHS 1256−1257B is included in the left panel of Figure 6. We note that none of these companions have (J − K)_MKO or J_MKO − W2 colors as red as CWISE J0506+0738. Of the remaining planetary-mass companions that lack 3–5 μm photometry, only 2M1207b (J − K = 3.07 ± 0.23 mag; Chauvin et al. 2004, 2005; Mohanty et al. 2007; Patience et al. 2010) and HD 206893B (J − K = 3.36 ± 0.08 mag; Delorme et al. 2017; Milli et al. 2017; Kammerer et al. 2021; Meshkat et al. 2021; Ward-Duong et al. 2021) have redder J − K colors than CWISE J0506+0738.

Young brown dwarfs have been shown to have enhanced photometric variability compared to field-age brown dwarfs (Biller et al. 2015; Metchev et al. 2015; Schneider et al. 2018; Vos et al. 2020, 2022a). Most brown dwarfs with detected variability at 3–5 μm, measured largely with Spitzer/IRAC, have amplitudes of a few percent or less (see compilation in Vos et al. 2020). Multiepoch photometry from WISE generally does not have the precision to detect such variability (Mace 2015; H. Brooks et al. 2023, in preparation). However, objects with extremely high-amplitude variability could be distinguished in multiepoch WISE data.

Given tentative evidence of near-infrared photometric variability (see Section 3.1), we investigated WISE (Wright et al. 2010) and NEOWISE (Mainzer et al. 2011, 2014) data for evidence of mid-infrared variability for CWISE J0506+0738. WISE/NEOWISE have been scanning the mid-infrared sky for over 10 yr, and a typical location on the sky has been observed with the W1 and W2 filters every six months since early 2010. ¹⁷ During each ∼1 day visit, 10–15 individual exposures are typically acquired. We chose to analyze these single exposures as opposed to epochal coadds (e.g., "unTimely"; Meisner et al. 2022) because CWISE J0506+0738 is brighter than the nominal threshold where single-exposure photometry becomes unreliable, especially at W2 (∼14.5 mag; Schneider et al. 2016a), and the concern that the coadded frames would dilute any traces of photometric variability. Such coadded photometry may prove useful for future investigations of long-term/long-period variability.

We gathered photometry from the WISE/NEOWISE Single-exposure Source Catalogs (WISE Team 2020a, 2020b, 2020c; NEOWISE Team 2020) for CWISE J0506+0738 and the same set of known L, T, and Y dwarfs shown in Figure 6. Collectively, these objects should have comparable levels of low-amplitude variability generally undetectable by WISE. For each source, we measured the average and standard deviations of both W1 and W2 magnitudes. We omit frames with qual_frame values equal to zero as these frames likely have contaminated flux measurements. Because single-exposure frames are subject to astronomical transients (e.g., cosmic ray hits, satellite streaks), we excluded 4σ outliers from the set of single-exposure photometry for each source. We also excluded sources that were either blended or contaminated (e.g., bright star halos, diffraction spikes).

Figure 7 compares mean and standard deviation values, which show clear trends in both W1 and W2 photometry. We immediately identify four objects with magnitudes between 12 and 14.5 that have photometric scatter above the 5%–95% confidence interval (≳2σ) in either W1 or W2.

Zoom In Zoom Out Reset image size

2MASS 21392676+0220226 (2MASS J2139+0220) is a T1.5 dwarf (Burgasser et al. 2006) that is well-known for its large-amplitude infrared variability. Radigan et al. (2012) monitored 2MASS J2139+0220 and found J-band variability with a peak-to-peak amplitude of ∼26%, which until recent observations of VHS 1256−1257B (Zhou et al. 2022) was the highest amplitude variability found for any brown dwarf. Since the Radigan et al. (2012) study, this object has been the subject of numerous variability investigations (Apai et al. 2013; Khandrika et al. 2013; Karalidi et al. 2015), with Yang et al. (2016) finding variability of 11%–12% in Spitzer/IRAC ch1 and ch2 photometry. The extreme variability of 2MASS J2139+0220 is attributed to variations in the thickness of silicate clouds (Apai et al. 2013; Karalidi et al. 2015; Vos et al. 2022b). This object has also been shown to have a nearly edge-on inclination (Vos et al. 2017) and is a kinematic member of the ∼200 Myr old Carina-Near moving group (Zhang et al. 2021).

WISE J052857.68+090104.4 ( WISE J0528+0901) is a clear W1 outlier, originally classified as a late-M giant by Thompson et al. (2013) but later reclassified as a very low-gravity L1 brown dwarf member of the ∼20 Myr 32 Orionis group (Burgasser et al. 2016). This planetary-mass object has an anomalous J − W2 color, suggestive of excess flux at 5 μm, although Burgasser et al. (2016) found no evidence of circumstellar material or cool companions. The source may also be variable in the W2 band, but its fainter magnitude here makes it less distinct than comparably bright L and T dwarfs. Nevertheless, these data suggest that WISE J0528+0901 has an unusually dusty and variable atmosphere, making it a compelling source for future photometric monitoring.

PSO J318.5338−22.8603 is a clear W2 outlier and an exceptionally red β Pic member that has been shown to have large-amplitude infrared variability in the infrared (Biller et al. 2015; Vos et al. 2019), with a peak-to-peak amplitude of 3.4% in Spitzer/IRAC ch2 photometry (Biller et al. 2018). Interestingly, PSO J318.5338−22.8603 is an outlier in W2 and not in W1, which may indicate cloud depth effects given that the W1 and W2 bands probe different depths in the atmosphere.

2MASSW J0310599+164816 (2MASS J0310+1648AB) is another W2 outlier and is an optically classified L8 (Kirkpatrick et al. 2000). This object is a resolved (02) ∼equal brightness binary (Stumpf et al. 2010) that shows evidence of high-amplitude variability in the near-infrared (Buenzli et al. 2014). While the variability observations were not long enough to determine a true amplitude or period, the brightening rate of ∼2% per hour was the largest measured in the sample. While there is no clear evidence of youth for 2MASS J0310+1648AB in the literature, this object was typed as L9.5 (sl. red) in Schneider et al. (2014). Further investigation of the potential youth and cloud properties of this object may be warranted.

CWISE J0506+0738 joins this group of variability outliers as one of very few objects with both W1 and W2 scatter outside the 16%–84% confidence interval of comparable-brightness L and T dwarfs. To estimate the amplitude of variability associated with these deviations, we fit tenth-order polynomials to the scatter versus magnitude trends in W1 and W2 and calculated rms values by finding the magnitude offset (in quadrature) for our outlying targets. Assuming sinusoidal variability, rms values can be converted to peak-to-peak amplitudes with a multiplicative factor of $2\sqrt{2}$. Using the 16%–84% confidence region as uncertainties for the predicted values from the polynomial fits, we find peak-to-peak variability on the order of 13% ± 1% for W1 and 12% ± 2% for W2 for 2MASS J2139+0220, which is generally consistent with results from Spitzer (Yang et al. 2016). For CWISE J0506+0738, we estimate 15% ± 5% variability for W1 and 23% ± 9% variability for W2. Variability at these levels would certainly be extraordinary; however, we caution that the relatively low precision of WISE/NEOWISE single-exposure measurements may inflate these results. Future photometric and/or spectroscopic monitoring would help to explore the variability properties of CWISE J0506+0738.

CWISE J0506+0738 is faint at optical wavelengths and was therefore undetected by the Gaia mission (Gaia Collaboration et al. 2022). The currently available astrometry for CWISE J0506+0738 is insufficient for a parallax measurement. Because CWISE J0506+0738 has such an unusually shaped spectrum, standard spectral type versus absolute magnitude relations for normal, field-age brown dwarfs are not applicable. There have been efforts to create relations between absolute magnitudes and spectral types for low-gravity brown dwarfs; however, these are typically valid for spectral types earlier than L7 (e.g., Faherty et al. 2016; Liu et al. 2016). Faherty et al. (2013) found absolute photometry of the young L5 dwarf 2MASS J03552337+1133437 was fainter than field L5 dwarfs at wavelengths shorter than ∼2.5 μm, and brighter at longer wavelengths. Schneider et al. (2016b) investigated other young, red L dwarfs with measured parallaxes and found that K-band photometry produced photometric distances that aligned well with parallactic distances. This trend was also noted in Filippazzo et al. (2015), Faherty et al. (2016), and Liu et al.(2016).

Here, we use nine young, free-floating brown dwarfs (Table 2) with measured parallaxes (Liu et al. 2016; Best et al. 2020; Kirkpatrick et al. 2021; Gaia Collaboration et al. 2022) to compare measured distances to photometric distances based on absolute magnitude-spectral type relations for J_MKO, K_MKO, W1, and W2 (Dupuy & Liu 2012; Kirkpatrick et al. 2021; Figure 8). Consistent with prior results, we find that K_MKO-band photometric distances (average offset Δd =−0.8 pc, scatter σ_d = 3.3 pc) are generally more accurate than J_MKO (Δd = −10 pc, σ_d = 5.1 pc), W1 (Δd = +2.6 pc, σ_d = 3.8 pc), or W2 (Δd = +4.5 pc, σ_d = 4.1 pc) photometric distances. To ensure these values are not biased, we also evaluated the fractional difference for each photometric band, defined as Δd/d_plx, and find that K-band photometric distances are typically within 5% for this sample, compared to 52%, 11%, and 20% for J_MKO, W1, and W2, respectively.

Zoom In Zoom Out Reset image size

Using the absolute magnitude-spectral type relation from Dupuy & Liu (2012), a spectral type of L9 ± 1, and its measured K_MKO photometry, we estimate a photometric distance of ${32}_{-3}^{+4}$ pc for CWISE J0506+0738. Again, given the exceptional nature of this source and its unknown multiplicity, we advise that this distance estimate be used with caution until it can be confirmed with a trigonometric parallax.

Young brown dwarfs are often associated both spatially and kinematically with young, nearby moving groups, thereby serving as invaluable age benchmarks.

To assess the potential moving group membership of CWISE J0506+0738, we use the BANYAN Σ algorithm (Gagné et al. 2018a), which deploys a Bayesian classifier to assign probabilities of moving group membership through 6D coordinate alignment (position and velocity) to 26 known moving groups in the solar neighborhood. We used the position and proper motion of CWISE J0506+0738 from UKIRT and UHS measurements (Table 1) and our measured RV from the NIRES spectrum (Section 4.3). With these values alone, we find an 82% membership probability in the β Pictoris moving group (BPMG; Zuckerman et al. 2001), a 3% membership probability in the AB Doradus moving group (ABDMG; Zuckerman et al. 2004), and a 15% probability of being unassociated with any moving group. The predicted/optimal distances for membership in BPMG and ABDMG are 32 pc and 64 pc, respectively; our estimated distance clearly aligns with the former. If we include the distance estimate in the BANYAN Σ algorithm, the probability of BPMG membership goes up to 99%.

We also tested the kinematic membership of CWISE J0506+0738 using the LACEwING analysis code (Riedel et al. 2017). Again, using just the position, proper motion, and RV of CWISE J0506+0738, we find nonzero probabilities for ABDMG (56%), the Argus Moving Group (71%), BPMG (28%), the Columba Association (52%), and the Tucana-Horologium Association (6%). Note that LACEwING is stricter in assigning membership probabilities than BANYAN, with bona fide BPMG members having a maximum membership probability of ∼70% when only proper motion and RV are used (Riedel et al. 2017). If we use our photometric distance as an additional constraint, BPMG is returned as the group with the highest probability of membership at 86%.

Membership in the BPMG is clearly favored for CWISE J0506+0738 although a directly measured distance is necessary for confirmation. If confirmed, CWISE J0506+0738 would have the latest spectral type and lowest mass among free-floating BPMG members, following PSO J318.5338−22.8603 (Liu et al. 2013a). Several candidate members with L7 or later spectral types have also been proposed (Best et al. 2015; Schneider et al. 2017; Kirkpatrick et al. 2021; Zhang et al. 2021; however, see Hsu et al. 2021). PSO J318.5338−22.8603 has proven to be an exceptionally valuable laboratory for studying planetary-mass object atmospheres (Biller et al. 2015; Allers et al. 2016; Faherty et al. 2016; Biller et al. 2018). A second planetary-mass object in this group that bridges the L/T transition will further contribute to these studies.

Assuming β Pic membership, we can use the group age of 22 ± 6 Myr (Shkolnik et al. 2017) to estimate the mass of CWISE J0506+0738. To do this, we must first estimate the luminosity (L_bol) or effective temperature (T_eff) of the source. For the former, we used the empirical K-band bolometric correction/spectral type relation for young brown dwarfs quantified in Filippazzo et al. (2015). Combining this with the UHS K-band magnitude and our distance estimate, we infer a bolometric luminosity of log(L_bol/L_⊙) = −4.55 ± 0.12. We caution that this value is based on our estimated distance from Section 5.3 and will need to be updated when a measured parallax becomes available. We then used the solar-metallicity evolutionary models of Marley et al. (2021) to infer a mass of 7 ± 2 M_Jup. The evolutionary models also provide a radius of 1.32 ± 0.03 R_Jup for these parameters, consistent with the radii of low-gravity late-type L dwarfs (Filippazzo et al. 2015). Combining this radius with our bolometric luminosity, we find T_eff = 1140 ± 80 K. This is ∼130 K cooler than a field-age L9 (Kirkpatrick et al. 2021), consistent with previous works showing low-gravity late-Ls tend to be ∼100–200 K cooler than field-age objects at the same spectral type (Filippazzo et al.2015; Faherty et al. 2016). In particular, this temperature is 50–100 K cooler than T_eff  estimates of PSO J318.5338−22.8603 (Liu et al. 2013a; Miles et al. 2018), consistent with the appearance of CH₄ absorption at lower temperatures.

The predicted mass of 7 ± 2 M_Jup is well below the deuterium-fusion minimum mass of 14 M_Jup commonly used to distinguish brown dwarfs from planetary-mass objects. As such, this object helps bridge the mass gap between the lowest mass free-floating β Pic members and directly imaged exoplanets, such as 51 Eri b (∼T6.5; Macintosh et al. 2015; Rajan et al. 2017). CWISE J0506+0738 could also help to constrain the effective temperature of the L/T transition at an age of ∼20–25 Myr (Binks & Jeffries 2014; Bell et al. 2015; Messina et al. 2016; Nielsen et al. 2016; Shkolnik et al. 2017; Miret-Roig et al. 2020). CWISE J0506+0738 would be one of the youngest objects to join a small but growing number of benchmark substellar objects with known ages at the L/T transition such as HD 203030B (30–150 Myr; Metchev & Hillenbrand 2006; Miles-Páez et al. 2017), 2MASS J13243553+6358281 (∼150 Myr; Looper et al. 2007; Gagné et al. 2018b), HIP 21152B, and other T-type Hyades members (∼650 Myr; Kuzuhara et al. 2022; Schneider et al. 2022), Indi Ba (∼3.5 Gyr; Scholz et al. 2003; Chen et al. 2022), and the white dwarf companion COCONUTS-1 (∼7 Gyr; Zhang et al.2020).