A Large Sample of JWST/NIRSpec Brown Dwarfs: New Distant Discoveries

TL;DR

The paper systematically mines 41,283 JWST/NIRSpec PRISM/CLEAR spectra to assemble 68 brown dwarfs, including 12 new distant candidates, with two objects exceeding 5 kpc—the most distant known in the Milky Way. It employs nested-sampling fits to three atmospheric grids (Sonora Elf Owl, LOWZ, SAND) to extract $T_ ext{eff}$, $ ext{log }g$, [M/H], C/O, and cloud properties, revealing that cloud-free models fail at the L/T transition while the SAND cloudy grid better replicates observations for metal-poor atmospheres. Distances are derived from radii and the flux-scale parameter $R^2/D^2$, with several new distant dwarfs likely belonging to the thick disk or halo; evolutionary-model-based parameters underscore tensions when $ ext{log }g$ is uncertain. A preliminary attempt to map the vertical metallicity gradient with these distant substellar objects finds no robust trend, constrained by sample size and model-systematics. Overall, the work substantially expands the census of distant brown dwarfs and emphasizes the importance of cloud physics and metal-poor atmospheres in interpreting ultracool spectra for Galactic structure studies.

Abstract

Brown dwarfs are essential probes of stellar and planetary formation, yet their low luminosities pose challenges for detection at large Galactic distances. The James Webb Space Telescope (JWST), with its unprecedented near-infrared sensitivity, enables the discovery and characterization of distant substellar objects, including those in the Milky Way's thick disk and halo. We conducted a systematic search using over 40,000 publicly available JWST/NIRSpec PRISM/CLEAR spectra and identified 68 brown dwarfs through spectral template matching and visual inspection. Among them, 12 are newly identified candidates, including 8 T dwarfs and 4 M/L dwarfs, most at distances exceeding 1 kpc. Remarkably, two sources -- JWST J001418.22-302223.2 and JWST J033240.07-274907.8 -- are found at distances greater than 5 kpc, making them the most distant brown dwarfs within the Milky Way. Spectral fits were performed using a nested sampling Monte Carlo algorithm with three model grids: Sonora Elf Owl, LOWZ, and SAND. The analysis reveals that cloud-free models are unable to reproduce L/T transition spectra, whereas the SAND model provides a more accurate representation of cloud effects in metal-poor environments. With the newly identified distant brown dwarfs, we also investigated the vertical metallicity gradient of brown dwarfs. Overall, the metallicities do not show an evident trend with Galactic height $|Z|$, due to the limited sample size and the uncertainties in metallicity measurements.

Figures (7)

• Figure 1: Flux calibration check for the two most distant brown dwarf candidates. Red points represent the NIRCam photometric measurements, while blue points show the synthetic photometry obtained by convolving the observed spectra with the corresponding NIRCam filter transmission curves.
• Figure 2: Comparison of the observed spectra of 2MASS J00470038$+$6803543 (dashed blue) with the best-fit parameters of cloud-free Sonora Elf Owl model (gray) and the cloudy SAND model (red).
• Figure 3: JWST/NIRSpec PRISM/CLEAR spectra of the brown dwarf candidates, shown as deep blue lines with light blue shaded regions indicating the uncertainties. The best-fit model spectra are overplotted in red. We show the object names and corresponding best-fit model parameters in the legend.
• Figure 4: (continued)
• Figure 5: The blue histogram shows the distance distribution of all brown dwarfs in our sample, while the red histogram highlights the distribution of twelve new brown dwarf candidates.