Disentangling disc and atmospheric signatures of young brown dwarfs with JWST/NIRSpec

TL;DR

This work tackles the challenge of separating disc and atmospheric signatures in young brown dwarfs by conducting joint atmospheric retrievals and disc modelling on JWST/NIRSpec spectra spanning 0.97–5.27 μm. The authors combine radiative transfer with line-by-line opacities, chemical equilibrium plus deviations, and a disc slab plus warm blackbody ring to extract temperature structures, abundances, isotope ratios, and disc properties for two TW Hydrae members. They detect more than twenty species, including 13CO and C18O, measure near-solar C/O ratios, and uncover warm disc material with inner cavities and hot CO gas, underscoring the necessity of coupled atmosphere–disc analyses. The results showcase JWST/NIRSpec's power to probe young substellar atmospheres and circumplanetary discs, providing benchmarks for planet formation models and guiding future high-resolution follow-up.

Abstract

Young brown dwarfs serve as analogues of giant planets and provide benchmarks for atmospheric and formation models. JWST has enabled access to near-infrared spectra of brown dwarfs with unprecedented sensitivity. We aim to constrain their chemical compositions, temperature structures, isotopic ratios, and disc emission. We perform retrievals and disc modelling on JWST/NIRSpec medium-resolution ($R \approx 2700$) spectra spanning 0.97--5.27 $μ$m, combining radiative transfer, line-by-line opacities, parameterised temperature profiles, and flexible equilibrium chemistry. We include a disc ring with blackbody continuum and optically thin CO emission. We detect over twenty species, including $^{12}$CO, H$_2$O, CO$_2$, SiO, and hydrides. The CO band at 4.6 $μ$m reveals $^{13}$CO and C$^{18}$O. Carbon isotope ratios are $^{12}$C/$^{13}$C = $79^{+14}_{-11}$ (TWA 27A) and $75^{+2}_{-2}$ (TWA 28); oxygen ratios are $^{16}$O/$^{18}$O = $645^{+80}_{-70}$ and $681^{+53}_{-50}$. Both objects show excess infrared emission, consistent with warm ($\approx 650$ K) blackbody rings, and optically thin CO from hot gas ($\geq 1600$ K) needed to match the red spectra. The atmospheric C/O ratios are $0.54 \pm 0.02$ (TWA 27A) and $0.59 \pm 0.02$ (TWA 28), consistent with solar values. We characterise the atmospheres and discs of two young brown dwarfs through joint constraints on temperature, composition, isotopes, and discs, demonstrating JWST/NIRSpec's ability to probe young objects and circumplanetary discs.

Figures (15)

• Figure 1: JWST/NIRSpec observations and atmospheric models. Full wavelength coverage of JWST/NIRSpec observations and model fits for TWA 27A and TWA 28. a shows observed spectra (black) and best-fit model spectra (coloured lines) for TWA 27A (blue) and TWA 28 (orange) across the full wavelength range. b displays relative residuals $\Delta F_{\lambda}/F_{\lambda}= (F_{\lambda,{\rm obs}}-F_{\lambda,{\rm model}})/F_{\lambda,{\rm obs}}$ of the best-fit model. c presents the same data in log-scale, with blackbody contribution indicated and coverage of each grating (G140H, G235H, G395H) shown as shaded regions. The increasing offset between datasets at redder wavelengths reflects distinct blackbody contributions in each system. d, e, and f show zoomed regions from each grating with key absorption features labelled.
• Figure 2: Atmospheric temperature profiles for TWA 27A (a) and TWA 28 (b). Left panels: Retrieved profiles from JWST/NIRSpec observations (G140H+G235H+G395H, blue lines) and excluding G140H (green lines), with 1-, 2-, and 3-$\sigma$ confidence intervals (shaded regions). CRIRES+ profile for TWA 28 gonzalezpicosESOSupJupSurvey2024 and a cloudless Sonora Diamondback model at $T_{\mathrm{eff}}=2400$ K, $\log{g}=4.0$morleySonoraSubstellarAtmosphere2024 are shown for comparison. Right panels: Residuals between median G140H+G235H+G395H profiles and other datasets.
• Figure 3: Molecular detections via cross-correlation analysis. CCF of selected molecules for TWA 27A and TWA 28, computed between model residuals (with and without each species) and template spectra over radial velocities up to 2000 km s$^{-1}$. S/N ratios are calculated by normalising the CCF peak to the standard deviation of the CCF-ACF difference (secondary panels). ACF is the autocorrelation function of each molecule.
• Figure 4: Chemical abundance offsets from solar-composition chemical equilibrium. The 1-, 2-, and 3-$\sigma$ confidence intervals of $\alpha$ as defined in \ref{['eq:alpha']} are shown for each of the retrieved species using chemical equilibrium models. The median value for each object is plotted as a horizontal dashed line.
• Figure 5: Corner plot of the posterior distributions of selected parameters for TWA 27A and TWA 28. The parameters are the radius, surface temperature, surface gravity, logarithm of the volume mixing ratio of the H$^{-}$ bound-free opacity, the $\alpha$-parameter of water, the carbon isotopologue ratio, the effective temperature and size of the blackbody, and the logarithm of the column density, the excitation temperature and the size of the slab model. The titles indicate the median and the 16th and 84th percentiles of the posterior distributions for each target retrieved from the fit to the entire wavelength range.