We present a spectrally and temporally resolved detection of the optical Mg I triplet at 7.8σ in the extended atmosphere of the ultra-hot Jupiter KELT-9 b, adding to the list of detected metal species in the hottest gas giant currently known. Constraints are placed on the density and radial extent of the excited hydrogen envelope using simultaneous observations of Hα and Hβ under the assumption of a spherically symmetric atmosphere. We find that planetary rotational broadening of v rot = 8.2 +0.6 −0.7 km s −1 is necessary to reproduce the Balmer line transmission profile shapes, where the model including rotation is strongly preferred over the non-rotating model using a Bayesian information criterion comparison. The time-series of both metal line and hydrogen absorption show remarkable structure, suggesting that the atmosphere observed during this transit is dynamic rather than static. We detect a relative emission feature near the end of the transit which exhibits a P-Cygni-like shape, evidence of material moving at ≈ 50 − 100 km s −1 away from the planet. We hypothesize that the in-transit variability and subsequent P-Cygni-like profiles are due to a flaring event that caused the atmosphere to expand, resulting in unbound material being accelerated to high speeds by stellar radiation pressure. Further spectroscopic transit observations will help establish the frequency of such events.

No abstract

No abstract

As followup to our recent detection of a pre-transit signal around HD 189733 b, we obtained full pre-transit phase coverage of a single planetary transit. The pre-transit signal is again detected in the Balmer lines but with variable strength and timing, suggesting that the bow shock geometry reported in our previous work does not describe the signal from the latest transit. We also demonstrate the use of the Ca II H and K residual core flux as a proxy for the stellar activity level throughout the transit. A moderate trend is found between the pre-transit absorption signal in the 2013 data and the Ca II H flux. This suggests that some of the 2013 pre-transit hydrogen absorption can be attributed to varying stellar activity levels. A very weak correlation is found between the Ca II H core flux and the Balmer line absorption in the 2015 transit, hinting at a smaller contribution from stellar activity compared to the 2013 transit. We simulate how varying stellar activity levels can produce changes in the Balmer line transmission spectra. These simulations show that the strength of the 2013 and 2015 pre-transit signals can be reproduced by stellar variability. If the pre-transit signature is attributed to circumplanetary material, its evolution in time can be described by accretion clumps spiraling towards the star, although this interpretation has serious limitations. Further high-cadence monitoring at Hα is necessary to distinguish between true absorption by transiting material and short-term variations in the stellar activity level.

No abstract