Evidence for stellar contamination in the transmission spectra of HAT-P-12b

Abstract

Transmission spectroscopy characterizes the wavelength dependence of transit depth, revealing atmospheric absorption features in planetary terminator regions. In this context, different optical transmission spectra of HAT-P-12b reported in previous studies exhibited discrepant atmospheric features (e.g., Rayleigh scattering, alkali absorption). We aim to understand the atmosphere of HAT-P-12b using two transit spectroscopic observations by the Gran Telescopio CANARIAS (GTC), and to search for evidence of stellar activity contaminating the transmission spectra, which might be the reason behind the discrepancies. We used Gaussian processes to account for systematic noise in the transit light curves and used nested sampling for Bayesian inferences. We performed joint atmospheric retrievals using the two transmission spectra obtained by GTC OSIRIS, as well as previously published results, coupled with stellar contamination corrections for different observations. The retrieved atmospheric model exhibits no alkali absorption signatures, but shows tentative molecular absorption features including $\rm H_2O$, $\rm CH_4$ and $\rm NH_3$. The joint retrieval of the combined additional public data analysis retrieves similar results, but with a higher metallicity. Based on Bayesian model comparison, the discrepancies of the transmission spectra of HAT-P-12b can be explained by the effect of different levels of unocculted stellar spots and faculae. In addition, we did not find strong evidence for a cloudy or hazy atmosphere from the joint analysis, which is inconsistent with prior studies based on the observations of Hubble Space Telescope.