JADES: An Abundance of Ultra-Distant T- and Y-Dwarfs in Deep Extragalactic Data

TL;DR

This study demonstrates that deep JWST/NIRCam imaging in extragalactic fields can reveal a rich population of ultra-cool brown dwarfs far beyond the solar neighborhood. The authors introduce NIFTY, a Bayesian fitting tool that compares JWST photometry to three atmosphere model grids to infer effective temperature, surface gravity, metallicity, and distances, using MIRI data up to ~15 μm to constrain the coldest sources. They identify 41 brown dwarfs and candidates in the JADES GOODS-S/N fields, including 10 Y-dwarfs (T_eff < 500 K) and 31 T-dwarfs (500–1200 K) out to 5–6 kpc, with most objects showing sub-solar metallicities and several displaying measurable proper motions. The work discusses the implications for Milky Way structure and brown dwarf demographics, and highlights contamination risks for ultra-high-redshift galaxy searches (e.g., Capotauro), underscoring the need for follow-up spectroscopy and expanded JWST surveys to refine the census and physics of these cold substellar populations.

Abstract

Ultra-cool T- (T$_{\mathrm{eff}} \approx$ 500 - 1200 K) and Y-dwarfs (T$_{\mathrm{eff}}$ $\lessapprox 500$ K) have historically been found only a few hundred parsecs from the Sun. The sensitivity and wavelength coverage of the NIRCam instrument on board the James Webb Space Telescope offer a unique method for finding low-temperature brown dwarfs in deep extragalactic datasets out to multiple kiloparsecs. Here we report on the selection of a sample of 41 brown dwarf and brown dwarf candidates across the JWST Advanced Deep Extragalactic Survey (JADES) in the GOODS-S and GOODS-N regions. We introduce a new open-source Bayesian tool, the Near-Infrared Fitting for T and Y-dwarfs (\texttt{NIFTY}), to derive effective temperatures, metallicities, and distances from JWST photometry. We find that 31 candidates have fits consistent with T-dwarf temperatures out to 5 - 6 kpc, and 10 candidates have fits consistent with Y-dwarf temperatures out to 1 - 2 kpc. The majority of the sources are best fit with sub-solar metallicity models, consistent with them being subdwarfs in the Milky Way thick disk and halo. We report proper motions for nine brown dwarf candidates (three are newly presented), and calculate the number density of T- and Y-dwarfs as a function of temperature and distance above the Milky Way midplane. We further discuss how Y-dwarfs can serve as contaminants in the search for ultra-high-redshift galaxies. Together, these results demonstrate the power of deep JWST extragalactic imaging to probe the coldest substellar populations far beyond the solar neighborhood, providing new constraints on the Milky Way's structure and brown dwarf demographics.

Figures (13)

• Figure 1: Color criteria used to select brown dwarf candidates in this work. In each panel, the colors of the underlying JADES GOODS-S population are represented with contours. Our 41 selected JADES brown dwarf candidates are shown using tan circles with black outlines. To show the colors spanned by brown dwarf models, we plot predicted color tracks derived from the Sonora Elf Owl models with downward triangles ([M/H] = +1.0), squares ([M/H] = 0.0), and upward triangles ([M/H] = -1.0), with the points colored by the T$_{\mathrm{eff}}$ values as shown in the color bar on the right. In each panel, we plot the color criteria used in the paper with black dashed lines. Left: F277W - F444W color plotted against F115W - F150W color. Middle: F277W - F444W color plotted against F410M - F444W color. There are sources outside the selection criteria in each panel as we only require selection using either the F277W - F444W vs F115W - F150W or the F277W - F444W vs F410M - F444W criteria. Right: F277W - F444W color plotted against F115W - F277W color, where we show, in a grey hashed region, an additional color criterion designed to reject LRDs. We show a sample of GOODS-S and GOODS-N LRDs from perezgonzalez2024, rinaldi2024, and kocevski2025 as small red points. To help exclude these LRDs while including each of our candidates, we suggest a refined selection criteria which we plot as a red dashed line.
• Figure 2: JADES v1.0 (Left) GOODS-S footprint and (Right) GOODS-N footprint over which we searched for brown dwarfs. We mark the full GOODS-S footprint including ancillary data with a grey line, and the JADES specific data with a blue line. Additionally, we plot the SMILES survey area with a light red line, and the additional area from the JADES MIRI parallels in F770W, F1280W, and F1500W with a dark red line. We plot the brown dwarfs and brown dwarf candidates recovered by our selection with tan points with black outlines. We note that the areas are not scaled proportionally between the GOODS-S and GOODS-N regions in each panel, and we indicate a 5$^{\prime}$ scale bar for reference in each panel.
• Figure 3: (Top right) Example NIFTY Fit to Source JADES-GS-BD-1. The JWST NIRCam and MIRI photometry is plotted in black, corresponding to the $2^{\prime\prime} \times 2^{\prime\prime}$ thumbnails below. The NIFTY median SED, and a 1$\sigma$ confidence interval, is plotted in red, with the median photometry in each filter shown with red squares. On the plot, we indicate the NIFTY parameters resulting from a fit with the Sonora Elf Owl models. (Left) The corner plot showing the posterior distributions for the free parameters from the Sonora Elf Owl fit to JADES-GS-BD-1.
• Figure 4: SED plots for the brown dwarf and brown dwarf candidates in our sample (black points in each panel), along with the Sonora Elf Owl fits and 1$\sigma$ confidence interval from NIFTY (red region), with the median NIFTY fit photometry plotted as red squares. In each panel, we indicate 2$\sigma$ upper limits with downward facing arrows. We also provide the Sonora Elf Owl model parameters in each panel, and below each SED we plot $2^{\prime\prime} \times 2^{\prime\prime}$ thumbnails for the majority of the filters that were used to observe these sources. This figure is continued in Figures \ref{['fig:SED_plots_pt2']}, \ref{['fig:SED_plots_pt3']}, and \ref{['fig:SED_plots_pt4']} in Appendix \ref{['sec:appendix_parameters']}.
• Figure 5: Best-fit effective temperature plotted against distance for the brown dwarf candidates selected in this work, derived using the Sonora Elf Owl models. The primary sample is plotted with black circles. In the background, we color the plot in three ranges, showing rough temperatures for L, T, and Y-dwarfs. Additionally, we plot the CEERS brown dwarf candidates presented in hainline2024a with grey squares, and the COSMOS-Web brown dwarf candidates presented in chen2025 with grey diamonds. For both of these samples, we re-fit the measured photometry from those studies with NIFTY using the Sonora Elf Owl models. To aid in understanding the distribution of sources, in dashed lines we plot the maximum distance allowable for Sonora Elf Owl models with F444W flux = 100 nJy, 20 nJy, 10 nJy, and 5 nJy. We are able to detect Y-dwarfs out to $\sim 2$kpc with JADES, demonstrating the power of deep NIRCam imaging.