Upper limits on atmospheric abundances of KELT-11b and WASP-69b from a retrieval approach

TL;DR

The paper tackles how to constrain the atmospheric composition of two low-density hot Jupiters, WASP-69 b and KELT-11 b, using high-resolution transmission spectroscopy. It combines cross-correlation with forward-model spectra and injection-recovery tests, plus atmospheric retrievals, to bound molecular abundances and cloud-deck pressuress. A tentative H$_2$O signal is reported for KELT-11 b, while H$_2$S and CH$_4$ remain undetected in both planets; WASP-69 b shows no robust H$_2$O detection under the data quality and processing choices. The two complementary approaches yield broadly consistent upper limits and highlight the role of high-altitude clouds in masking spectral features, illustrating a robust framework for future high-resolution atmospheric retrievals of exoplanets.

Abstract

WASP-69b and KELT-11b are two low-density hot Jupiters, which are expected to show strong atmospheric features in their transmission spectra. Such features offer valuable insights into the chemical composition, thermal structure, and cloud properties of exoplanet atmospheres. High-resolution spectroscopic observations can be used to study the line-forming regions in exoplanet atmospheres and potentially detect signals despite the presence of clouds. We aimed to detect various molecular species and constrain the chemical abundances and cloud deck pressures using high-resolution spectroscopy. We observed multiple transits of these planets with CARMENES and applied the cross-correlation method to detect atmospheric signatures. Further, we used an injection-recovery approach and retrievals to place constraints on the atmospheric properties. We detected a tentative H$_2$O signal for KELT-11b but not for WASP-69b, and searches for other molecules such as H$_2$S and CH$_4$ resulted in non-detections for both planets. By investigating the signal strength of injected synthetic models, we constrained which atmospheric abundances and cloud deck pressures are consistent with our cross-correlation results. In addition, we show that a retrieval-based approach leads to similar constraints of these parameters.

Figures (17)

• Figure 1: Data reduction steps of a representative wavelength range for night 1 of WASP-69 b. Top panel: Unprocessed spectra as produced by the CARACAL pipeline. Middle panel: After normalisation to bring all spectra onto a common continuum level, and masking of deep telluric lines. Bottom panel: After the removal of stellar and telluric lines with SYSREM.
• Figure 2: Model spectra (top) and $K_\mathrm{p}$-$v_\mathrm{offset}$ maps for the combination of all nights (bottom) of H$_2$O, H$_2$S, and CH$_4$ for KELT-11 b. The grey shaded areas show the wavelength regions of the spectral orders used for the analysis. The dotted lines indicate the expected location of the planetary signal.
• Figure 3: Same as Fig. \ref{['Fig_Models+Detmaps_K11']} but for WASP-69 b.
• Figure 4: Results of the S/N grid analysis for KELT-11 b. Shown are the retrieved S/N values of injected synthetic models with varying molecular abundances and cloud deck pressures (in bar) for H$_2$O, H$_2$S, and CH$_4$ (note the different colour scales). The dashed lines indicate the thresholds corresponding to a S/N of 1 (green), 3 (teal), and 5 (purple). Notably, the S/N for CH$_4$ is consistently below 3, so only the S/N=1 threshold is shown.
• Figure 5: Same as Fig. \ref{['Fig_SNR_Grid_K11']} but for WASP-69 b.