Chromium hydride in the atmosphere of WASP-31b: Evidence from HST and prospects for JWST

Marrick Braam 1,2,3 , Floris van der Tak 3,4 , Katy Chubb 5 , Michiel Min 5

  • 1 School of Geosciences, University of Edinburgh, Edinburgh
  • 2 Institute of Astronomy, KU Leuven, Leuven
  • 3 Kapteyn Astronomical Institute, University of Groningen, Groningen
  • 4 SRON Netherlands Institute for Space Research, Groningen
  • 5 SRON Netherlands Institute for Space Research, Utrecht

Abstract

A recent retrieval study indicated evidence for chromium hydride (CrH) in the atmosphere of WASP-31b, presenting the first evidence for the signatures of CrH in a hot Jupiter atmosphere. I will present the reanalysis of publicly available transmission data and the implications of the presence of CrH. Using the Bayesian retrieval framework TauREx II, we found evidence for the spectroscopic signatures of CrH, H2O and K on WASP-31b. Compared to a flat model without any signatures, a CrH-only model is preferred with a statistical significance of ~3.9σ. A model consisting of both CrH+H2O is found with ~2.6 and ~3σ confidence over a CrH-only and H2O-only model respectively. Furthermore, evidence for a Rayleigh scattering slope was found, indicating the presence of aerosols. At a retrieved temperature of ~1481 K, the atmosphere of WASP-31b is hot enough to host gaseous Cr-bearing species and the retrieved abundances agree well with predictions from thermal equilibrium chemistry and the findings in an L-type brown dwarf atmosphere. The detection and abundance of metal compounds have potential consequences for planet formation, atmospheric chemistry and cloud formation. The proximity of WASP-31b's temperature to the condensation temperature of Cr might imply a contrast between the evening and morning terminators. Additional retrievals using ground-based VLT FORS2 data lead to a non-detection of CrH. Future facilities such as JWST have the potential to confirm the detection and/or discover other CrH features.

Context

Chemical detections

The characterisation of exoplanet atmospheres is showing a wide diversity in compositions. So far, chemical detections include H2O, CH4 and CO2, but also exotic metallic compounds such as AlO, TiO and VO. This is no surprise, as hot Jupiters have the appropriate temperatures to host metal compounds in the gaseous phase, and these should be detectable through transmission spectroscopy. Also expected are chromium hydride (CrH) and iron hydride (FeH), the latter already has a few tentative detections and atomic chromium has also been found in hot Jupiter atmospheres. 

The source

The hot Jupiter WASP-31 b orbits an F-type star and was discovered by Anderson et al. in 2011. A low density and equilibrium temperature of 1393 K makes it a suitable candidate for atmospheric characterisation and put it in the right temperature range for metal hydrides. Using combined HST and Spitzer data, detections of the signatures of H2O, NH3, K and clouds/hazes have been reported, but disagreement exists between different analyses. Here, we use the TauREx retrieval framework to conduct a reanalysis of the publicly available data to search for the signatures of exotic species.

Discussion

CrH in atmospheres

Earlier work also found evidence for the presence of H2O and K, though later studies using ground-based data questioned the detection of K. Our results present the first evidence for CrH in a hot Jupiter atmosphere. At the retrieved Tatm~1481 K, chromium (Cr) is expected to be in gaseous phase with CrH as one of the main bearers, according to predictions from thermal equilibrium chemistry. The detection can have implications for cloud formation and give clues about the planet's formation history.

A possible CrH contrast?

Depending on the pressure, Cr(g) is expected to condense out between 1400-1520 K, close to the retrieved temperature of WASP-31b. Informed by the huge day-night temperature contrast of hot Jupiters, we might expect a difference between the ingress and egress spectrum of WASP-31b. Assuming the ingress spectrum to be fed by winds from the nightside, we would expect CrH to be condensed out, whereas the egress spectrum should show CrH in gaseous form. The simulated JWST observations in Figure 3 show the potential observability of this contrast in CrH.

Figure 3: Simulated JWST observations for just the ingress and egress of WASP-31b, focussed on the wavelength region relevant for CrH, created using TauREx Forward Models and PandExo. Retrieved model fits including H2O+CrH and H2O-only are also plotted.

Methodology

Retrieval Setup

TauREx II computes the radiative transfer of stellar light through the planetary atmosphere based on a range of parameters and finds the best fit to the observed spectrum through the MultiNest algorithm. Chemical constituents have a specific interaction with stellar light, described by the molecular and atomic opacities as taken from the ExoMol, HITEMP and MoLLIST databases. Furthermore, the opacities from collision-induced absorption, Rayleigh scattering and an optically thick cloud deck are included. In total, up to 28 free parameters can be retrieved. With the limited number of data points and sparse spectral coverage, only a small number of these parameters will statistically be required.

Quantifying Chemical Detections

Therefore, we followed a bottom-up approach by increasing the model complexity in steps. Comparison between two models is done by using their Bayesian Evidence E to calculate the Detection Significance (DS):

{DS}=\ln({E}_2)-\ln({E}_1)

Where model 2 has an extra parameter as compared to model 1. A DS greater than one provides evidence in favour of the more complex model, which can be translated into frequentist values using Jeffrey's scale. The different stages are:

  • The simplest atmospheric model lacking chemical signatures or a 'flat' spectrum.
  • Add the abundance of a specific chemical species as a free parameter, one by one.
  • The inclusion of H2O and another species, can be compared to both earlier stages.
  • Testing the addition of alkali metals and some high-Evidence species of previous stages.

In total, the spectrum of WASP-31b was retrieved assuming 53 different models.

Results

Model comparison

The Bayesian evidence of the 53 atmospheric models is shown in Figure 1, separated in vertical space according to the four different stages. Comparing the models can be done using the scale bars on the lower right. The second stage shows the models with a single atmospheric species, indicating the highest evidence for the presence of CrH. One stage higher, the inclusion of H2O leads to a further increase in evidence and weak evidence is found for potassium (K) in the final stage. A model including H2O and CrH is preferred with a confidence level of 4.4σ over the flat model. Compared to a CrH- or H2O-only model, the confidence corresponds to 2.6 and 3.0σ, respectively. 

Figure 1: Model comparison for WASP-31b using the Bayesian evidence for 53 atmospheric models. The flat model is shown as the orange dot, and cyan dots indicate higher complexity in the form of a chemical species, as labelled. One stage higher, blue dots represent a model that includes H2O and an additional parameter. For the final stage, the strongest evidence model from lower complexities is complemented by K, Na, and some of the more likely species of lower stages. The horizontal scale bars indicate the statistical preference for a more complex model.

Atmospheric model WASP-31b

Hence, out of the models fitted in this study, the spectrum of WASP-31b is best represented by a model that includes H2O, CrH and K in addition to H2, He, a grey cloud deck and Rayleigh scattering. The retrieval results in an atmospheric temperature of ~1481 K. The atmospheric model and the observed transmission spectrum (with HST and Spitzer) of WASP-31b are shown in the left panel of Figure 2. The right panel shows the individual contributions of atmospheric constituents to the opacity. The characteristic features of H2O in the near-infrared (1.0, 1.2 and 1.4 μm; blue line) and K in the visible (0.77 μm; dark red line) are easily recognised, as are the continuum opacities resulting from a grey cloud, Rayleigh scattering and collision-induced absorption (CIA). The presence of CrH results in six absorption signatures between 0.7 and 1.5 μm, as shown by the orange line, of which the features around 0.88 and 0.77 μm are driving the detection.

Figure 2: Transmission spectrum of WASP-31b, focussing on the wavelengths covered by HST STIS and WFC3. The left panel shows the best-fitting atmospheric model and the right panel shows individual opacity contributions of atmospheric constituents. The vertical error bars indicate the observed transit depths, and the different shadings in the upper panel represent 1 and 2σ regions.

Conclusions

From Combined HST STIS and WFC3 and Spitzer data, we find that the spectrum WASP-31b is best explained by a model including H2O, CrH, K (and H2, He, a grey cloud, CIA and Rayleigh scatter). This represents the first evidence for CrH features in a hot Jupiter atmosphere. CrH has implications for cloud formation, atmospheric chemistry and atmospheric dynamics. Informed by these dynamics, a CrH contrast can be expected in WASP-31b, and there is a possibility of detection by egress/ingress spectroscopy using JWST.

References

Publication: Braam, M., van der Tak, F. F. S., Chubb, K. L., & Min, M. 2021, A&A, 646, A17.
Discovery WASP-31b: Anderson, D. R., Collier Cameron, A., Hellier, C., et al. 2011, A&A, 531, A60.
TauREx II: Waldmann, I. P., Tinetti, G., Rocchetto, M., et al. 2015, ApJ, 802, 107.
ExoMol: Tennyson, J., Yurchenko, S. N., Al-Refaie, A. F., et al. 2016, J. Mol. Spectr., 327, 73
HITEMP: Rothman, L. S., Gordon, I. E., Barber, R. J., et al. 2010, J. Quant. Spectr. Rad. Transf., 111, 2139
MoLLIST: Bernath, P. F. 2020, J. Quant. Spectr. Rad. Transf., 240, 106687