We present an improved, hybrid CPU-GPU atmospheric retrieval code, Helios-r2, which is applicable to medium-resolution emission spectra of brown dwarfs, in preparation for precision atmospheric spectroscopy in the era of the James Webb Space Telescope. The model is available as open-source code on the Exoclimes Simulation Platform. We subject Helios-r2 to a battery of tests of varying difficulty. The simplest test involves a mock retrieval on a forward model generated using the same radiative transfer technique, the same implementation of opacities, and the same chemistry model. The least trivial test involves a mock retrieval on synthetic spectra from the Sonora model grid, which uses a different radiative transfer technique, a different implementation of opacities, and a different chemistry model. A calibration factor, which is included to capture uncertainties in the brown dwarf radius, distance to the brown dwarf and flux calibration of the spectrum, may compensate, sometimes erroneously, for discrepancies in modeling choices and implementation. We analyze spectra of the benchmark brown dwarf GJ 570 D and the binary brown dwarf companions in the Epsilon Indi system. The retrieved surface gravities are consistent with previous studies and/or values inferred from dynamical masses (for Epsilon Indi Ba and Bb only). There remains no clear criterion on how to reject unphysical values of the retrieved brown dwarf radii. The inferred radii and corresponding masses should be taken with great caution. The retrieved carbon-to-oxygen ratios and metallicity depend on whether chemical equilibrium is assumed.

Self-consistent model grids of brown dwarfs involve complex physics and chemistry, and are often computed using proprietary computer codes, making it challenging to identify the reasons for discrepancies between model and data as well as between the models produced by different research groups. In the current study, we demonstrate a novel method for analyzing brown dwarf spectra, which combines the use of the Sonora, AMES-Cond and HELIOS model grids with the supervised machine learning method of the random forest. Besides performing atmospheric retrieval, the random forest enables information content analysis of the three model grids as a natural outcome of the method, both individually on each grid and by comparing the grids against one another, via computing large suites of mock retrievals. Our analysis reveals that the different choices made in modelling the alkali line shapes hinder the use of the alkali lines as gravity indicators. Nevertheless, the spectrum longward of 1.2 micron encodes enough information on the surface gravity to allow its inference from retrieval. Temperature may be accurately and precisely inferred independent of the choice of model grid, but not the surface gravity. We apply random forest retrieval to three objects: the benchmark T7.5 brown dwarf GJ 570D; and Epsilon Indi Ba (T1.5 brown dwarf) and Bb (T6 brown dwarf), which are part of a binary system and have measured dynamical masses. For GJ 570D, the inferred effective temperature and surface gravity are consistent with previous studies. For Epsilon Indi Ba and Bb, the inferred surface gravities are broadly consistent with the values informed by the dynamical masses.

The chemical composition of an exoplanet is a key ingredient in constraining its formation history. Iron is the most abundant transition metal, but has never been directly detected in an exoplanet due to its highly refractory nature. KELT-9b (HD 195689b) is the archetype of the class of ultra-hot Jupiters that straddle the transition between stars and gas-giant exoplanets and serve as distinctive laboratories for studying atmospheric chemistry, because of its high equilibrium temperature of 4050 +/- 180 K. These properties imply that its atmosphere is a tightly constrained chemical system that is expected to be nearly in chemical equilibrium and cloud-free. It was previously predicted that the spectral lines of iron will be detectable in the visible range of wavelengths. At these high temperatures, iron and several other transition metals are not sequestered in molecules or cloud particles and exist solely in their atomic forms. Here, we report the direct detection of atomic neutral and singly-ionized iron (Fe and Fe+), and singly-ionized titanium (Ti+) in KELT-9b via the cross-correlation technique applied to high-resolution spectra obtained during the primary transit of the exoplanet.

We analyze the emission spectrum of the hot Jupiter WASP-12b using our HELIOS-R retrieval code and HELIOS-K opacity calculator. When interpreting Hubble and Spitzer data, the retrieval outcomes are found to be prior-dominated. When the prior distributions of the molecular abundances are assumed to be log-uniform, the volume mixing ratio of HCN is found to be implausibly high. A VULCAN chemical kinetics model of WASP-12b suggests that chemical equilibrium is a reasonable assumption even when atmospheric mixing is implausibly rigorous. Guided by (exo)planet formation theory, we set Gaussian priors on the elemental abundances of carbon, oxygen and nitrogen with the Gaussian peaks being centered on the measured C/H, O/H and N/H values of the star. By enforcing chemical equilibrium, we find substellar O/H and stellar to slightly superstellar C/H for the dayside atmosphere of WASP-12b. The superstellar carbon-to-oxygen ratio is just above unity, regardless of whether clouds are included in the retrieval analysis, consistent with Madhusudhan et al. (2011). Furthermore, whether a temperature inversion exists in the atmosphere depends on one's assumption for the Gaussian width of the priors. Our retrieved posterior distributions are consistent with the formation of WASP-12b in a solar-composition protoplanetary disk, beyond the water iceline, via gravitational instability or pebble accretion (without core erosion) and migration inwards to its present orbital location via a disk-free mechanism, and are inconsistent with both in-situ formation and core accretion with disk migration, as predicted by Madhusudhan et al. (2017). We predict that the interpretation of James Webb Space Telescope WASP-12b data will not be prior-dominated.

We present an open-source retrieval code named HELIOS-Retrieval (hereafter HELIOS-R), designed to obtain chemical abundances and temperature-pressure profiles from inverting the measured spectra of exoplanetary atmospheres. In the current implementation, we use an exact solution of the radiative transfer equation, in the pure absorption limit, in our forward model, which allows us to analytically integrate over all of the outgoing rays (instead of performing Gaussian quadrature). Two chemistry models are considered: unconstrained chemistry (where the mixing ratios are treated as free parameters) and equilibrium chemistry (enforced via analytical formulae, where only the elemental abundances are free parameters). The nested sampling algorithm allows us to formally implement Occam's Razor based on a comparison of the Bayesian evidence between models. We perform a retrieval analysis on the measured spectra of the HR 8799b, c, d and e directly imaged exoplanets. Chemical equilibrium is disfavored by the Bayesian evidence for HR 8799b, c and d. We find supersolar C/O, C/H and O/H values for the outer HR 8799b and c exoplanets, while the inner HR 8799d and e exoplanets have substellar C/O, substellar C/H and superstellar O/H values. If these retrieved properties are representative of the bulk compositions of the exoplanets, then they are inconsistent with formation via gravitational instability (without late-time accretion) and consistent with a core accretion scenario in which late-time accretion of ices occurred differently for the inner and outer exoplanets. For HR 8799e, we find that spectroscopy in the K band is crucial for constraining C/O and C/H. HELIOS-R is publicly available as part of the Exoclimes Simulation Platform (ESP; www.exoclime.org).

Optical phase curves have become one of the common probes of exoplanetary atmospheres, but the information they encode has not been fully elucidated. Building on a diverse body of work, we upgrade the Flexible Modeling System (FMS) to include scattering in the two-stream, dual-band approximation and generate plausible, three-dimensional structures of irradiated atmospheres to study the radiative effects of aerosols or condensates. In the optical, we treat the scattering of starlight using a generalisation of Beer's law that allows for a finite Bond albedo to be prescribed. In the infrared, we implement the two-stream solutions and include scattering via an infrared scattering parameter. We present a suite of four-parameter general circulation models for Kepler-7b and demonstrate that its climatology is expected to be robust to variations in optical and infrared scattering. The westward and eastward shifts of the optical and infrared phase curves, respectively, are shown to be robust outcomes of the simulations. Assuming micron-sized particles and a simplified treatment of local brightness, we further show that the peak offset of the optical phase curve is sensitive to the composition of the aerosols or condensates. However, to within the measurement uncertainties, we cannot distinguish between aerosols made of silicates (enstatite or forsterite), iron, corundum or titanium oxide, based on a comparison to the measured peak offset ($41^\circ \pm 12^\circ$) of the optical phase curve of Kepler-7b. Measuring high-precision optical phase curves will provide important constraints on the atmospheres of cloudy exoplanets and reduce degeneracies in interpreting their infrared spectra.