2022ApJ...931...86H

The atmospheric structure of WASP-12b has been hotly contested for years, with disagreements on the presence of a thermal inversion as well as the carbon-to-oxygen ratio, C/O, due to retrieved abundances of H₂O, CO₂, and other included species such as HCN and C₂H₂. Previously, these difficult-to-diagnose discrepancies have been attributed to model differences; assumptions in these models were thought to drive retrievals toward different answers. Here, we show that some of these differences are independent of model assumptions and are instead due to subtle differences in the inputs, such as the eclipse depths and line-list databases. We replicate previously published retrievals and find that the retrieved results are data driven and are mostly unaffected by the addition of species such as HCN and C₂H₂. We also propose a new physically motivated model that takes into consideration the formation of H^– via the thermal dissociation of H₂O and H₂ at the temperatures reached in the dayside atmosphere of WASP-12b, but the data's current resolution does not support its inclusion in the atmospheric model. This study raises the concern that other exoplanet retrievals may be similarly sensitive to slight changes in the input data.