Table of Contents
Fetching ...

Observation of undepleted phosphine in the atmosphere of a low-temperature brown dwarf

Adam J. Burgasser, Eileen C. Gonzales, Samuel A. Beiler, Channon Visscher, Ben Burningham, Gregory N. Mace, Jacqueline K. Faherty, Zenghua Zhang, Clara Sousa-Silva, Nicolas Lodieu, Stanimir A. Metchev, Aaron Meisner, Michael Cushing, Adam C. Schneider, Genaro Suarez, Chih-Chun Hsu, Roman Gerasimov, Christian Aganze, Christopher A. Theissen

Abstract

The atmospheres of low-temperature brown dwarfs and gas giant planets are expected to contain the phosphine molecule, PH$_3$ However, previous observations have shown much lower abundances of this molecule than predicted by atmospheric chemistry models. We report JWST spectroscopic observations of phosphine in the atmosphere of the brown dwarf Wolf 1130C. Multiple absorption lines due to phosphine are detected around 4.3 $μ$m, from which we calculate a phosphine abundance of 0.100$\pm$0.009 parts per million. This abundance is consistent with disequilibrium atmospheric chemistry models that reproduce the phosphine abundances in Jupiter and Saturn, and is much higher than abundances previously reported for other brown dwarfs or exoplanets.

Observation of undepleted phosphine in the atmosphere of a low-temperature brown dwarf

Abstract

The atmospheres of low-temperature brown dwarfs and gas giant planets are expected to contain the phosphine molecule, PH However, previous observations have shown much lower abundances of this molecule than predicted by atmospheric chemistry models. We report JWST spectroscopic observations of phosphine in the atmosphere of the brown dwarf Wolf 1130C. Multiple absorption lines due to phosphine are detected around 4.3 m, from which we calculate a phosphine abundance of 0.1000.009 parts per million. This abundance is consistent with disequilibrium atmospheric chemistry models that reproduce the phosphine abundances in Jupiter and Saturn, and is much higher than abundances previously reported for other brown dwarfs or exoplanets.

Paper Structure

This paper contains 21 sections, 3 equations, 10 figures, 2 tables.

Figures (10)

  • Figure 1: JWST NIRSpec spectra of Wolf 1130C compared to ULAS J1029+0935 and models.
  • Figure 2: Sonora Elf-Owl models fitted to the spectrum of Wolf 1130C. (A) The observed spectrum fitted by models with artificially suppressed PH$_3$ abundances (EOwl 2024ApJ...963...73M); (B) the same but using models with PH$_3$ abundances set by vertical mixing (EOwl+ 2024ApJ...973...60B). Both figures compare the absolute flux-calibrated spectrum (black lines) to the best-fitting model (magenta line) and 20 draws from the MCMC posterior chains (pale magenta lines) as an indication of the model uncertainty. Below each panel we show the difference (observed minus computed, or O-C) between the observed and best-fitting model spectra and the $\pm$5$\sigma$ uncertainty of the former (grey bands). The dashed black lines indicate zero flux.
  • Figure 3: Retrieval analysis of the NIRSpec/G395H spectrum of Wolf 1130C. (A-F) The G395H spectrum (black line) compared to the median likelihood atmosphere model from our retrieval analysis (mustard line). The two are nearly indistinguishable, with a reduced $\chi^2_r$ = 6. Single pixel deviations at 5.06 $\upmu$m and 5.14 $\upmu$m are due to noise. (G) Retrieved abundances (solid vertical lines) for H$_2$O (blue), CH$_4$ (orange), CO (pink), CO$_2$ (red), NH$_3$ (brown), H$_2$S (cyan), and PH$_3$ (green) plotted as a function of logarithmic pressure (log P) in the photosphere. Shaded regions around each line indicate the $\pm$1$\sigma$ uncertainties on the retrieved abundances from the retrieval posterior distributions (Fig. \ref{['fig:retrieval3']}). Dashed lines are the predicted abundances for a [M/H] = $-$0.7, C/O = 0.26 thermochemical equilibrium model. CO$_2$ is not visible in this plot as its equilibrium abundance is $<$10$^{-12}$ (Fig. \ref{['fig:chemistry2']}) Grey shading indicates the approximate pressure region of the photosphere (visible layer of the atmosphere) in the 3 to 5 $\upmu$m region. The retrieved abundances are assumed to be constant at all pressures.
  • Figure 4: Chemical equilibrium model predictions of PH$_3$ abundances in the atmosphere of Wolf 1130C. Each panel displays the equilibrium mole fraction abundances for PH$_3$ in green to purple shading as a function of temperature and pressure. Abundance values are indicated by the associated color bars, and isoabundances are indicated in 1 dex (dotted) and 5 dex (solid) increments (black on green, white on purple), with values between 0.1 ppb (parts per billion) through 0.1 ppm (parts per million) labeled. The near-vertical black dashed line indicates the condensation curve for NH$_4$H$_2$PO$_4$, which depletes PH$_3$ at lower temperatures. The three black curves indicate the median and $\pm$1$\sigma$ uncertainty boundaries for the atmospheric thermal profile of Wolf 1130C derived from our retrieval analysis (Fig. \ref{['fig:retrieval2']}). (A) A model with solar metallicity and nominal phosphorus chemistry. (B) Same as panel A, but for one tenth of solar metallicity. (C) Same as panel B, but assuming a more exothermic enthalpy of formation for P$_4$O$_6$, leading to depletion of PH$_3$ at warmer temperatures supp.
  • Figure S1: Posterior probability distributions for Sonora Elf-Owl models fitted to the spectrum of Wolf 1130C. Results are shown for fits based on models with PH$_3$ abundances set by vertical mixing (EOwl+ 2024ApJ...973...60B). Panels along the diagonal show histograms of the marginalized posterior distributions for model parameters T$_\mathrm{eff}$, $\log_\mathrm{10}{\mathrm{g}}$, [M/H], C/O, $\log\kappa_{zz}$, and radius $R$. Off-diagonal panels display the marginalized posterior distributions among parameter pairs as contour plots, with contours set at 25%, 50% and 75% confidence intervals. Solid magenta lines in the posterior distributions indicate the best-fitting value of each parameter, while dashed black lines in the diagonal panels indicate quantiles of 16%, 50% and 84% of the posterior distributions, listed in Table \ref{['tab:fit_results']}. This figure was generated using the corner.py package corner.
  • ...and 5 more figures