JWST 1.5 μm and 4.8 μm Photometry of Y Dwarfs

The target list of Y dwarfs was selected from the literature as available in 2020. We selected spectroscopically confirmed brown dwarfs as listed in J. D. Kirkpatrick et al. (2019, their Table 11) with T_eff < 500 K within 15 pc of the Sun. For the NIRCam program (PID 2473), brown dwarfs calculated to be too close to known background objects during Cycle 1 were removed from the sample, as were known binary brown dwarfs. Two targets in that sample were observed for principal investigator Albert as part of GTO programs (PIDs 1189 and 1230) using NIRISS in imaging mode, which included two additional targets in the 15 pc Y-dwarfs sample (WISE J085510.83–071442.5 and WISEPA J182831.08+265037.8) as well as an additional point-spread function (PSF) calibration target, UGPS J072227.51–054031.2, a nearby T9 brown dwarf. The full sample is listed in Table 1 with discovery references, effective temperature values, parallax measurements, and CatWISE2020 W2 magnitudes. Hereafter, the targets will be written as WISE-hhmm, e.g., WISE-0855.

Table 1. Sample of Brown Dwarfs

Short Name	Full Name	Disk.	Teff	Teff	Parallax	W2
		Ref.	Est. K	Ref.	(mas)	(mag)
NIRCam F150W+F480M Survey (Program ID 2473)
WISE-0304	WISE J030449.03–270508.3	(1)	465  ±  88	(a)	73.1  ±  2.6	15.578  ±  0.041
WISE-0336AB	WISE J033605.05–014350.4AB	(2)	460  ±  79	(b)	99.8  ±  2.1	14.664  ±  0.023
WISE-0359	WISE J035934.06–540154.6	(2)	$44{3}_{-19}^{+23}$	(c)	73.6  ±  2.0	15.412  ±  0.026
WISE-0410	WISEPA J041022.71+150248.5	(3)	451  ±  88	(a)	151.3  ±  2.0	14.104  ±  0.017
WISE-0535	WISE J053516.80–750024.9	(2)	$49{6}_{-23}^{+28}$	(c)	68.7  ±  2.0	14.996  ±  0.020
WISE-0647	WISE J064723.23–623235.5	(4)	393  ±  88	(a)	99.5  ±  1.7	15.115  ±  0.018
WISE-0713	WISE J071322.55–291751.9	(2)	464  ±  88	(a)	109.3  ±  2.1	14.327  ±  0.016
WISE-0734	WISE J073444.02–715744.0	(2)	$46{6}_{-20}^{+24}$	(c)	74.5  ±  1.7	15.241  ±  0.022
WISE-0825	WISE J082507.35+280548.5	(5)	$38{7}_{-15}^{+15}$	(c)	152.6  ±  2.0	14.655  ±  0.024
WISE-0830	WISEA J083011.95+283716.0	(6)	367  ±  79	(a)	90.6  ±  13.7a	16.004  ±  0.072
WISE-1141	WISEA J114156.67–332635.5	(7)	460  ±  79	(a)	104.0  ±  2.9	14.632  ±  0.030
WISE-1206	WISE J120604.38+840110.6	(8)	$47{2}_{-20}^{+26}$	(c)	84.7  ±  2.1	15.173  ±  0.023
WISE-1405	WISEPC J140518.40+553421.4	(3)	$39{2}_{-15}^{+16}$	(c)	158.2  ±  2.6	14.098  ±  0.014
WISE-1446	CWISEP J144606.62–231717.8	(9)	$35{1}_{-13}^{+16}$	(c)	103.8  ±  5.0	15.955  ±  0.072
WISE-1541	WISEPA J154151.66–225025.2	(3)	$41{1}_{-17}^{+18}$	(c)	166.9  ±  2.0	14.218  ±  0.030
WISE-1639	WISEA J163932.75+184049.4	(9)	511  ±  79	(a)	61.9  ±  4.7	15.481  ±  0.036
WISE-1738	WISEPA J173835.53+273258.9	(3)	450  ±  88	(a)	130.9  ±  2.1	14.519  ±  0.017
WISE-2056	WISEPC J205628.90+145953.3	(3)	$48{1}_{-20}^{+26}$	(c)	140.8  ±  2.0	13.925  ±  0.016
WISE-2220	WISE J222055.31–362817.4	(2)	452  ±  88	(a)	95.5  ±  2.1	14.807  ±  0.024
WISE-2354	WISEA J235402.79+024014.1	(8)	$34{7}_{-11}^{+13}$	(c)	130.6  ±  3.3	15.018  ±  0.030
NIRISS F480M Imaging (Program ID 1230)
UGPS-0722	UGPS J072227.51–054031.2	(10)	569  ±  45	(a)	242.8  ±  2.4	12.198  ±  0.008
WISE-0855	WISE J085510.83–071442.5	(11)	250  ±  50	(a)	439.0  ±  2.4	13.820  ±  0.029
NIRISS F480M Imaging (Program ID 1189)
WISE-1828	WISEPA J182831.08+265037.8	(3)	406  ±  88	(a)	100.3  ±  2.0	14.393  ±  0.016

Notes. Discovery references are as follows: (1) D. J. Pinfield et al. (2014), (2) J. D. Kirkpatrick et al. (2012), (3) M. C. Cushing et al. (2011), (4) J. D. Kirkpatrick et al. (2013), (5) R. L. Griffith et al. (2012), (6) D. C. Bardalez Gagliuffi et al. (2020), (7) C. G. Tinney et al. (2014), (8) A. C. Schneider et al. (2015), (9) F. Marocco et al. (2020), (10) P. W. Lucas et al. (2010), (11) K. L. Luhman (2014).T_eff references are as follows: (a) J. D. Kirkpatrick et al. (2024), (b) J. D. Kirkpatrick et al. (2021), (c) S. A. Beiler et al. (2024). Parallaxes are from J. D. Kirkpatrick et al. (2021) except for WISE-1446, where the parallax is from S. A. Beiler et al. (2024). The W2 magnitudes are proper-motion-corrected values (w2mpro_pm) from the CatWISE2020 Catalog (F. Marocco et al. 2021), curated by IRSA. ^aUpdated astrometry for WISE-0830 is given in Table 4.

Download table as:  ASCIITypeset image

NIRCam enables dual-band imaging in the blue (λ ≤ 2.4 μm) and red (λ ≥ 2.5 μm) channels simultaneously (M. J. Rieke et al. 2023). For the blue channel, we selected the F150W (1.5 μm) wide-band filter (Δλ/λ = 10%) because most Y dwarfs had measurements in the ground-based H-band filter which is centered near 1.6 μm (see J. D. Kirkpatrick et al. 2019, 2021). For the red channel, we selected a filter at a wavelength where the Y dwarfs emit significant flux, the F480M (4.8 μm) medium-band filter (Δλ/λ = 5%; C. V. Morley et al. 2014; A. C. Schneider et al. 2015).

We observed the sample in five groups of four targets, and specified that observations within a group were to be executed within 3 days. This strategy was designed to minimize any wave-front drift and, for each target, to be able to use the three other targets as PSF reference stars (assuming they were not binaries).

The integration times ranged, for a given target, between 30 minutes and 2.5 hr, which provided a signal-to-noise ratio (SNR) > 1000 in F480M. Achieving such a high SNR (or roughly 10⁷ photons) at 4.8 μm was crucial to be able to use the kernel-phase interferometry analysis technique for our binary search. The SNRs in F150W were much lower (10–100) because of the extremely red H − W2 colors of ≥6 magnitudes for this sample of Y dwarfs. Note that this high SNR is considered separately from the absolute flux calibration uncertainty for NIRCam, currently listed at 0.9% (F480M) and 0.4% (F150W) in the respective photom reference files found on the JWST Calibration Reference Data System (CRDS).²²

Each target was observed over the $2\mathop{.}\limits^{^{\prime} }2\times 2\mathop{.}\limits^{^{\prime} }2$ field of view of module B in FULL readout mode with dithered exposures. The first 12 targets used a five-point subpixel dither type (SMALL-GRID-DITHER) that was found to be suboptimal for bad pixel correction. It was switched to a larger (2$\mathop{.}\limits^{^{\prime\prime} }$9-8$\mathop{.}\limits^{^{\prime\prime} }$9) five-point primary dither strategy (INTRAMODULEBOX) enabling better bad pixel interpolation for the remaining eight targets. We attempted to center the target near pixels 1200, 700 on the long-wave detector (F480M band) corresponding to pixels near 400, 1400 in the short-wave detector NIRCB2 (F150W band) by using a target offset of ${\rm{\Delta }}\alpha ,{\rm{\Delta }}\delta =+22\mathop{.}\limits^{^{\prime\prime} }0,-8\mathop{.}\limits^{^{\prime\prime} }0$. Errors in the position epoch entered in the Astronomer Proposal Tool caused significant deviations, in one case resulting in having the target off the F150W channel for two dithers. Table 2 lists the observations for targets observed as part of PID 2473.

Table 2. F150W and F480M Photometry of Our Y-dwarf Sample

Short Name	Observing Date	Int. Time	F150W	F480M
	Mid Exposure (UT)	(minutes)	(Vega mag)	(Vega mag)
WISE-0304	2022-09-23T21:58:02	89.5	21.872  ±  0.010	15.428  ±  0.005
WISE-0336AB	2022-09-22T13:13:10	40.3	22.038  ±  0.012	14.538  ±  0.005
WISE-0336Aa	2022-09-22T13:13:10	40.3	22.114  ±  0.016	14.727${}_{-0.023}^{+0.054}$
WISE-0336Ba	2022-09-22T13:13:10	40.3	24.938${}_{-0.105}^{+0.175}$	16.534${}_{-0.252}^{+0.120}$
WISE-0359	2022-09-23T17:51:38	89.5	22.281  ±  0.010	15.268  ±  0.005
WISE-0410	2022-09-23T23:43:56	26.8	20.167  ±  0.010	13.999  ±  0.005
WISE-0535	2022-06-27T22:27:30	62.6	22.866  ±  0.013	14.795  ±  0.005
WISE-0647	2022-11-11T16:38:29	58.2	23.332  ±  0.013	14.919  ±  0.005
WISE-0713	2022-11-12T08:42:33	31.3	20.445  ±  0.010	14.183  ±  0.005
UGPS-0722b	2023-04-17T22:33:44	18.9	17.26  ±  0.03	12.016  ±  0.011
WISE-0734	2022-06-28T00:07:35	76.1	21.706  ±  0.011	15.162  ±  0.004
WISE-0825	2022-11-12T10:10:45	40.3	22.831  ±  0.014	14.418  ±  0.005
WISE-0830	2022-11-12T22:49:59	120.8	25.261  ±  0.026	15.644  ±  0.005
WISE-0855b	2022-11-19T09:25:29	36.5	23.98  ±  0.10	13.679  ±  0.011
WISE-1141	2023-06-29T22:27:31	40.3	20.580  ±  0.010	14.582  ±  0.005
WISE-1206	2023-02-27T21:31:43	71.6	21.227  ±  0.010	15.112  ±  0.005
WISE-1405	2023-03-02T12:43:40	25.1	21.756  ±  0.012	13.930  ±  0.005
WISE-1446	2023-03-02T09:15:03	143.2	23.672  ±  0.013	15.694  ±  0.005
WISE-1541	2023-03-02T11:18:04	28.6	22.004  ±  0.011	14.018  ±  0.005
WISE-1639	2023-06-28T21:47:19	13.4	21.316  ±  0.014	13.487  ±  0.005
WISE-1738	2022-06-29T17:21:58	31.3	20.365  ±  0.010	14.464  ±  0.005
WISE-1828b	2022-07-28T17:52:53	35.8	23.07  ±  0.10	14.242  ±  0.011
WISE-2056	2023-06-28T09:03:58	22.4	20.057  ±  0.001	13.762  ±  0.005
WISE-2220	2023-07-01T21:31:28	49.2	21.161  ±  0.011	14.745  ±  0.005
WISE-2354	2022-06-29T19:09:18	58.2	23.098  ±  0.013	14.930  ±  0.005

Notes. ^aThe individual F150W and F480M magnitudes for the WISE-0336 system were determined from the aperture photometry for the unresolved system and the flux ratios for each component measured by P. Calissendorff et al. (2023). ^bThe F150W magnitudes for UGPS-0722, WISE-0855, and WISE-1828 were synthesized from spectroscopic observations by P. W. Lucas et al. (2010), K. L. Luhman et al. (2024), and M. C. Cushing et al. (2021). Uncertainties for these values were estimated to be equal to the spectral flux calibration uncertainty.

Download table as:  ASCIITypeset image

The target list was supplemented with three additional brown dwarfs, UGPS-0722, WISE-0855, and WISE-1828, observed with NIRISS in imaging mode using filters F480M + CLEARP to search for companions (see Table 2). One of the sources, UGPS-J0722, is a late T dwarf and not a Y dwarf, and is warmer than the rest of the sample, with T_eff ≈ 540 K (e.g., S. K. Leggett et al. 2021).

The WISE-0855 observation was accompanied by a contemporaneous calibrator star observation of UGPS-0722, intended to serve as a PSF reference for both WISE-0855 and WISE-1828. During the first epoch of observations of WISE-0855 and UGPS-0722 the Fine Guidance Sensor lost the guide-star lock and repeat observations of both targets were performed. A large SNR (>1000) was achieved for all three targets. Observations were undithered (staring) and centered on a bad-pixel-free region of the NIRISS detector. All observations with NIRISS were obtained in FULL subarray mode, except UGPS-0722, which was observed in SUB80 (80 × 80 pixels) to prevent saturation.

We downloaded the stacked images and associated catalogs constructed from the five dithered exposures (i2d.fits) processed by the Data Management System (DMS, version 1.9.6; H. Bushouse et al. 2022). The photometry catalogs produced by the default DMS level 3 pipeline contained anomalous magnitude entries for a few targets, which prompted us to revisit this step by performing our own aperture photometry using the photutils package (version 1.8.0; L. Bradley et al. 2023). For each filter (F480M/F150W), we first determined the centroid of the PSF by fitting a quadratic function in the 3 × 3 pixels around the pixel with peak intensity, then estimated the local sky level from statistics in an annulus with radii of 4.92/6.082 and 7.083/9.496 pixels around the PSF center and subtracted this level. Next, we performed aperture photometry by summing pixel flux within a radius of 3.757/3.199 pixels, defined in the CRDS reference file, jwst_nircam_apcorr_0004.fits, as the radius encompassing 70% of the PSF-encircled energy. Finally, we applied an aperture correction by multiplying the extracted flux by 1.4863/1.4485. This yielded flux measurements, F, expressed in megajansky/steradian that we converted to Vega magnitudes by applying the mag_AB − mag_Vega = −3.85/ − 2.15 mag offset (reference: jwst_nircam_abstovegaoffset_0002.fits) from the AB magnitude:

$$\begin{eqnarray}&&{\mathrm{mag}}_{\mathrm{AB}}=-2.5\,{\mathrm{log}}_{10}\left(\displaystyle \frac{F[\mathrm{MJy}/\mathrm{sr}]\times T\times A[\mathrm{sr}/\mathrm{pixel}]}{3631[\mathrm{Jy}]\times 1{0}^{-6}[\mathrm{MJy}/\mathrm{Jy}]}\right),\end{eqnarray} \tag{ 1 }$$

where T is the throughput calibration, PHOTMJSR = 1.456 ± 0.006/2.338 ± 0.021, A is the steradian-to-pixel area conversion, and PIXAR_SR = 9.332 × 10⁻¹⁴/2.287 × 10⁻¹⁴, found in the jwst_nircam_photom_0153/0150.fits reference file.

The measurement uncertainty considers the readout noise and photon noise given as a pixel map by the DMS pipeline (the ERR extension of the cal.fits) but also includes the uncertainty in subtracting the background level. These combine to very small statistical errors of order 0.08% (F480M) and 0.2%–2% (F150W). The absolute flux calibration uncertainty is added in quadrature and, with a precision of 0.4% in the F480M, dominates the error budget for the photometric measurements reported in Table 2. For the F150W filter, the absolute calibration uncertainty, 0.9%, is comparable to the statistical noise.

We present updated photometry for both components of the known binary WISE-0336A/B from that reported in P. Calissendorff et al. (2023). We treated the A+B components as unresolved to obtain the aperture photometry of the system in exactly the same manner as for the other NIRCam targets. We then used the binary contrast published in P. Calissendorff et al. (2023) of ${\rm{\Delta }}{\rm{F}}480{\rm{M}}={\rm{F}}480{{\rm{M}}}_{{\rm{B}}}-{\rm{F}}480{{\rm{M}}}_{{\rm{A}}}=1.8{1}_{-0.31}^{+0.14}$ and ${\rm{\Delta }}{\rm{F}}150{\rm{W}}={\rm{F}}150{{\rm{W}}}_{{\rm{B}}}-F150{{\rm{W}}}_{{\rm{A}}}=2.8{2}_{-0.11}^{+0.19}$ to calculate the magnitude of each component separately. The uncertainties were propagated using a Monte Carlo simulation of 5 × 10⁵ realizations. Results are resilient to aperture size selection.

We reprocessed the data with the DMS (version 1.12.5), enabling the charge_migration step, up to the end of level 2, which produced cal.fits images. Then we ran a custom 1/f correction on the three data sets observed in FULL mode (both WISE-0855 and WISE-1828 observations) while using directly the cal.fits for UGPS-0722 because the 1/f imprint is less pronounced on the SUB80 images. Using the custom 1/f-corrected data changed the photometry by less than 1% compared to using the cal.fits directly.

To extract photometry of NIRISS F480M images, we adopted the same procedure as performed on the NIRCam data sets and used by the DMS pipeline. We used the prescribed aperture and sky annulus radii of 4.68, 7.80, and 11.74 pixels, respectively, while the aperture correction was 1.458 (70% encircled energy), found in jwst_niriss_apcorr_0008.fits. The AB to Vega magnitude offset was mag_AB − mag_Vega = –3.4268, the PHOTMJSR = 1.220 ± 0.013, and the PIXAR_SR = 1.009 × 10⁻¹³ (jwst_niriss_abstovegaoffset_0003.fits, jwst_niriss_photom_0043.fits). The photometry has very small statistical errors of order 0.07%. The absolute flux calibration uncertainty is added in quadrature and, with a precision of 1.04%, dominates the error budget for the F480M photometric measurements reported in Table 2.

In order to compare the colors of these three NIRISS sources to the rest of our sample, we synthesized F150W photometry from available near-infrared spectra (P. W. Lucas et al. 2010; K. L. Luhman 2014; M. C. Cushing et al. 2021). The JWST user documentation supplies filter transmissions which include the instrument system throughput for each camera; we used the F150W NIRCam transmission for consistency with the other measurements. For reference, we explored the differences between the NIRCam and NIRISS photometric systems by synthesizing F150W and F480M magnitudes for each camera for five Y dwarfs with JWST spectra across the bandpasses (S. A. Beiler et al. 2024). The five Y dwarfs had a range in T_eff of 350–500 K, and the differences in the photometry were small, with no trend seen with T_eff. We found an average difference for NIRCam – NIRISS of ΔF150W = −0.040 ± 0.012 and ΔF480M = –0.010 ± 0.001 mag. Any systematic difference due to the different F480M throughput for the two cameras is therefore similar in size to the absolute flux calibration uncertainty.

We compare the photometry to synthetic photometry calculated by four cool brown dwarf models, each based on different atmosphere prescriptions. As a baseline model, the Sonora Bobcat models assume equilibrium chemistry (M. S. Marley et al. 2021). But, given the evidence for out-of-equilibrium chemistry in the atmosphere of ultracool dwarfs, the Sonora Elf Owl models include disequilibrium chemistry (S. Mukherjee et al. 2024), the ATMO2020++ models use disequilibrium chemistry and introduce adjustments to the adiabat profile (S. K. Leggett & P. Tremblin 2023), while B. Lacy & A. Burrows (2023) introduce water clouds in their disequilibrium models because the atmospheres of the coldest Y dwarfs, those with T_eff ≲ 400 K, are cold enough for water to condense (e.g., C. V. Morley et al. 2014). We note that the Sonora Cholla (T. Karalidi et al. 2021) and Diamondback (C. V. Morley et al. 2024) models do not include atmospheres as cold as those of the Y dwarfs.

All model sequences in Figure 1 are cloud-free except for the B. Lacy & A. Burrows (2023) sequence shown in the bottom-right panel. Solar and subsolar metallicities are included. Surface gravities of log g = 4.0 cm s⁻² and 4.5 cm s⁻² are shown. From evolutionary models we estimate that field dwarfs with ages of 2 to 4 Gyr have log g ≈ 4.5 cm s⁻² for 325 ≲ T_eff K ≲ 550 and log g ≈ 4.0 cm s⁻² for 200 ≲ T_eff K ≲ 325 (M. S. Marley et al. 2021). Model sequences' line types indicate metallicity and gravity, as shown in the legend in the top-right panel.

Figure 1 indicates that the ATMO2020++ (M. W. Phillips et al. 2020; S. K. Leggett et al. 2021; A. M. Meisner et al. 2023; S. K. Leggett & P. Tremblin 2025) and the B. Lacy & A. Burrows (2023) models reproduce the observed colors better than the Sonora Bobcat or Elf Owl models (M. S. Marley et al. 2021; S. Mukherjee et al. 2024). The T_eff values from the ATMO2020++ and the B. Lacy & A. Burrows (2023) models are also in good agreement with the values measured by S. A. Beiler et al. (2024) from the observed luminosities, while the Bobcat and Elf Owl models are not.

For brown dwarfs warmer than 350 K, the Sonora Bobcat F150W – F480M colors appear to be too red, while the Elf Owl colors appear to be too blue (assuming that this local sample of brown dwarfs is not unusually metal-poor). The likely cause of the red Bobcat color is the exclusion of disequilibrium chemistry, which increases the abundance of CO and CO₂ (e.g., K. J. Zahnle & M. S. Marley 2014) that absorb light at λ ≈ 4.7 μm and 4.2 μm, respectively (e.g., Z. Tu et al. 2024, their Figure 5), thereby decreasing the F480M flux. The likely cause of the blue Elf Owl color is an excess of near-infrared flux at F150W; it seems that standard radiative-convective atmospheres for cold brown dwarfs produce an excess of energy in the near-infrared (e.g., S. K. Leggett et al. 2017, 2021; T. Karalidi et al. 2021), which is the issue that the ATMO2020++ models address empirically by changing the pressure–temperature profile such that the deeper regions, where the near-infrared energy emerges, is colder. Although empirical, the adjustment is physically motivated by thermal changes due to rapid rotation and/or chemical changes in the deep layers due to condensation (e.g., S. K. Leggett et al. 2021; S. K. Leggett & P. Tremblin 2025).

The ATMO2020++ sequence does appear to show a systematic offset in color. This may be partly due to the definition of the F150W bandpass. For cold brown dwarfs the signal through this filter comes only from the narrow H-band flux peak at the extreme red of the bandpass. Hence, the color is sensitive to the definition of the red cutoff of the filter.

The ATMO2020++ and the B. Lacy & A. Burrows (2023) models calculate that the F480M absolute magnitude is a good indicator of T_eff for 275–600 K brown dwarfs because of its relatively small dependence on metallicity and gravity (ΔM_F480M ∼ 0.3 mag for Δlog g or Δ[m/H] ∼ 0.5 dex); compare the y-axis locations of the plus symbols in Figure 1. On the other hand, these models calculate that the F150W – F480M color is sensitive to both gravity and metallicity (Δ(F150W − F480M) ∼ 1.0 mag for Δlog g or Δ[m/H] ∼ 0.5 dex); compare the x-axis locations of the plus symbols in Figure 1.

WISE-0535 and WISE-1828 appear to be superluminous and/or redder than the other dwarfs in Figure 1. WISE-0830 may be a colder example of this group. We discuss each of these brown dwarfs individually below, in R.A. order.

A preliminary analysis of the spectral distribution of WISE-0535 using ATMO2020++ synthetic spectra by S. K. Leggett & P. Tremblin (2024) indicates that its superluminosity is due to multiplicity. We address this further in our companion paper (T. Vandal et al. 2025, in preparation).

This work and Matuszewska et al. (2025, in preparation) present the first near-infrared detection of the cold brown dwarf WISE-0830. Its location in Figure 1 suggests that it is a typical field dwarf with T_eff ≈ 350 K. However, WISE-0830 may be superluminous, depending on the (currently poorly defined) color–magnitude trend for the coldest objects. Superluminosity suggests lower metallicity, higher gravity (older age), or multiplicity. WISE-0830 has the highest tangential velocity in our sample (108 km s⁻¹; D. C. Bardalez Gagliuffi et al. 2020), possibly supporting an older age and lower metallicity for this Y dwarf compared to typical field dwarfs.

The extreme colors of WISE-1828 may indicate that it is multiple, has a high gravity, or is metal-poor. Recent studies using JWST imaging data find no evidence of a companion at separations >0.5 au (M. De Furio et al. 2023). Recent spectral analyses of JWST NIRSpec and MIRI data by B. W. P. Lew et al. (2024) and D. Barrado et al. (2023) find that the metallicity is approximately solar, and that the luminosity is consistent with evolutionary models if the system is a tight binary—an order-of-magnitude estimate by B. W. P. Lew et al. (2024), based on the radial velocity, suggests a separation of 20 Jupiter radii. S. K. Leggett & P. Tremblin (2025) find a good fit to the JWST NIRSpec and MIRI data with ATMO2020++ solar-metallicity models if the system is an equal-mass binary with unusually high gravity, i.e., a relatively massive and old system.

CWISEP J104756.81+545741.6 (hereafter WISE-1047, identified as "1047" in Figure 1) is not in our imaging sample, but is in the spectroscopic sample of S. A. Beiler et al. (2024), who synthesize JWST colors. S. A. Beiler et al. (2024) also provide a trigonometric parallax measurement for WISE-1047, whose discovery is presented in A. M. Meisner et al. (2020). S. A. Beiler et al. (2024) measure a parallax of 68.1 ± 4.9 mas, and calculate apparent magnitudes from their spectra of F150W = 22.10 and F480M = 16.33; the NIRSpec absolute calibration uncertainty is estimated to be less than 3% or 0.03 mag.²³

Figure 1 shows that WISE-1047 is blue or subluminous. Z. Tu et al. (2024) point out that the JWST spectra for this Y dwarf show unusually strong CO₂ and CO absorption (see their Figure 4). Atmospheric models calculate that decreasing gravity results in increasing CO and CO₂ absorption, and that these features are less sensitive to changes in metallicity (T. Karalidi et al. 2021; B. Lacy & A. Burrows 2023; S. K. Leggett & P. Tremblin 2025). The bandpass of the F480M filter directly samples the CO absorption, making it a useful indicator of the strength of this feature, an effect which could be masked by a broader filter.

Figure 1 suggests log g ≪ 4.0 and T_eff ≈ 400 K for WISE-1047. Evolutionary models (e.g., M. S. Marley et al. 2021) then suggest a mass less than 3 M_Jup and an age less than 0.2 Gyr for this Y dwarf. Taking the proper motion for this source from J. D. Kirkpatrick et al. (2021) and the parallax from S. A. Beiler et al. (2024), the BANYAN Σ tool (J. Gagné et al. 2018) calculates a 52% likelihood that WISE-1047 is a member of the Argus association, which has an age of ∼40 Myr (B. Zuckerman 2018). If this brown dwarf is indeed this young, then its mass is only 1 M_Jup and log g ≈ 3.3. Further improvement to the distance measurement and a radial velocity measurement would be useful for this object.

Water clouds are first expected to impact the photosphere when brown dwarfs cool to T_eff ∼ 350 K (e.g., A. Burrows et al. 2003; C. V. Morley et al. 2014; B. Lacy & A. Burrows 2023). The B. Lacy & A. Burrows (2023) disequilibrium-chemistry sequences shown in Figure 1 are for clear atmospheres and for atmospheres with thin water clouds at pressures of 0.4 bar with a particle size of 10 μm (E10 type). The addition of clouds makes F150W brighter and F480M fainter (B. Lacy & A. Burrows 2023), resulting in the bluer F150W – F480M colors seen in Figure 1.

The coldest objects in our sample are WISE-0336B and WISE-0855, with T_eff ≲ 300 K. The WISE-0336 binary components are separated by $0\mathop{.}\limits^{^{\prime\prime} }09$, or 1 au (P. Calissendorff et al. 2023). The reanalysis of the photometry for the system presented here produces colors within 0.07 magnitudes of the P. Calissendorff et al. (2023) values. Figure 1 suggests that WISE-0336A is a typical field Y dwarf, with T_eff between 400 and 450 K, and that WISE-0336B is slightly warmer than WISE-0855 but colder than the other Y dwarfs in the sample, with T_eff between 275 and 300 K. If the system has an age of ∼2 Gyr (e.g., W. M. J. Best et al. 2024), then the component masses are approximately 12 and 5 M_Jup (M. S. Marley et al. 2021).

Interestingly, WISE-0336B is significantly bluer than WISE-0855. Given that WISE-0336A appears to have a gravity and metallicity typical of the field, the bluer color for 0336B is unlikely to be due to an unusually low gravity or high metallicity. Instead, the B. Lacy & A. Burrows (2023) models suggest the difference is either that the atmosphere of WISE-0336B is cloudy and that of WISE-0855 is clear, or that WISE-0855 is cloudy but has a significantly higher gravity and/or lower metallicity than WISE-0336B. Recent analyses of JWST data for WISE-0855 support the former scenario, as H. Kühnle et al. (2024) find no evidence of water clouds, and the metallicity and gravity appears typical of the field (K. L. Luhman et al. 2024; H. Kühnle et al. 2024). The detection of water clouds may also be dependent on the surface gravity of the brown dwarf or on the viewing angle: Observations of L dwarfs suggest that the vertical extent of the water clouds is dependent on gravity (G. Suárez et al. 2023; G. Suárez & S. Metchev 2023), and the structured nature of surface storms and clouds results in colors being dependent on viewing angle (J. M. Vos et al. 2017).