Benchmark Brown Dwarfs as Chemical Laboratories: Linking System Bulk Properties to Atmospheric Retrievals

We present the first atmospheric retrieval analysis of two compositional benchmark mid-L dwarfs – SDSSJ141659.78 + 500626.4 and GJ 499 C – using the brewster retrieval framework, where the wide benchmark nature of these systems provides independent constraints on age and bulk composition from their stellar primaries. These targets were observed with James Webb Space Telescope using NIRSpec (Near Infrared Spectrograph) Prism and MIRI (Mid-Infrared Instrument) LRS (Low Resolution Spectroscopy), providing low-resolution (R (Formula presented) 100) spectra between 0.6–14.0 (Formula presented) m with high signal-to-noise ratios (SNR (Formula presented) 100–600). For SDSSJ141659.78 + 500626.4, the retrieved cloud combination of a high-altitude enstatite slab and low-altitude iron deck clouds matches phase-equilibrium predictions based on the primary star’s Mg/Si ratio. We retrieve a super-solar C/O = 0.71(Formula presented) and slightly metal-rich [M/H] = 0.22(Formula presented). This C/O ratio can only be reconciled with the value inferred for the primary if additional oxygen sequestration beyond the retrieved cloud mass is present, or if there are uncertainties in the adopted opacities or other model deficiencies. The inferred [C/H] and [O/H] are 0.31(Formula presented) 0.03 and 0.19(Formula presented) 0.03, respectively, which are consistent within the relatively large uncertainties of the host star abundances. For GJ 499 C, the retrieved silicon-monoxide and forsterite slab clouds are difficult to explain with simple phase-equilibrium assumptions, yielding Mg/Si (Formula presented) 1.9. We estimate C/O = 0.69(Formula presented) and [M/H] = 0.13(Formula presented). For both objects, the inferred radii of 0.85(Formula presented) (Formula presented) and 1.00(Formula presented) (Formula presented) and masses of 73.7(Formula presented) (Formula presented) and 68.0(Formula presented) (Formula presented) are consistent with evolutionary models and system ages, highlighting the plausibility of our results.