Exploring challenges of observing exoplanet spectra at high resolution

Thesis

In recent years, exoplanet spectroscopy has become a hot topic of research, towards the characterisation of other planets and fuelled by the existential question of extraterrestrial life. Ground-based high-resolution spectroscopy, whereby individual planetary spectral lines can be detected, affords unique characterisation opportunities. Compared to other astrophysical fields, high-resolution spectroscopy for the detection of exoplanet atmospheres is in its infancy, since it was only established a decade ago. Yet, the high-resolution cross-correlation spectroscopy technique has yielded tremendous insight into hot Jupiter and warm Saturn atmospheres. Despite its success, there are challenges that remain, which should be addressed if we are to progress towards more demanding targets. In this thesis I explore a number of such challenges.

First, I investigated the near-infrared transmission spectrum of the young sub-Neptune AU Mic b, analysing high-resolution spectra with standard cross-correlation techniques. Though the real data yielded null detections, I demonstrate that the data would be sensitive to a variety of atmospheric scenarios. I thus place a number of constraints on the atmospheric content of AU Mic b, with a highlight conclusion being that the atmosphere is unlikely to be in chemical equilibrium.

Then, I investigated the impact of stellar heterogeneities on high-resolution analyses. In this simulation-based study, I forward-modelled a number of stellar spectral time-series considering stellar rotation and magnetic phenomena, and compared the recovery of planet atmospheres. I found that the Rossiter-Mclaughlin effect can completely obscure planetary atmosphere detections - it is thus advisable to correct for this phenomenon explicitly in high-resolution analyses. I also investigated the impact of unocculted and occulted photospheric spots, concluding that occulted spots are likely to pose more of a danger in high-resolution cross-correlation spectroscopy.

Finally, I developed an alternative methodology to cross-correlation technique, namely the employment of Gaussian process regression to sequentially model the distinct spectral components. This bypasses cross-correlation and the associated issues such as spurious noise features in the cross-correlation maps, and statistical interpretation of a detection signal-to-noise. I applied this technique to archival high-resolution spectral data of canonical hot Jupiters HD 189733 b and 51 Pegasi b. While the Gaussian process framework was not sensitive to the real signals, previously detected at 5-sigma significance with cross-correlation analyses, it was sensitive to slightly amplified signals.

Authors: Meech, A

• DOI: 10.5287/ora-mz51qagad
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: exoplanets; stars; astrophysics; atmosphere; spectroscopy
• Subjects: Extrasolar planets