Using Doppler tomography for the characterisation of exoplanet atmospheres

  • Shannon Matthews

Student thesis: Doctoral ThesisDoctor of Philosophy


Characterising the atmospheres of exoplanets allows for the understanding of their formation, climate and habitability. There are three broad methods used for characterising the atmospheres of exoplanets: direct imaging, transmission spectroscopy, and (high-resolution) emission spectroscopy. The first, direct imaging, is currently only applicable to very large, hot planets orbiting far from their host star. Transmission spectroscopy, on the other-hand, is limited to planets that have a transit that is visible from the Earth, and is currently amenable to short-period large planets. Finally, high-resolution Doppler spectroscopy provides an avenue to study both transiting and non-transiting planets and also enables the radial-velocity of the planet to be directly measured, thereby allowing the absolute mass of non-transiting planets to be determined. This powerful method has also yielded some of the most robust atmosphere detections to date.

Currently, high-resolution Doppler spectroscopy studies detect atmospheric signals by cross-correlating observed data with a model atmospheric spectrum. This technique has been successful in detecting various molecular and atomic species such as H2O, CO, TiO and Fe. Here we present an alternative method of performing high-resolution Doppler spectroscopy, using Doppler tomography. We present several successful confirmations of chemical species from a range of exoplanets and provide a novel detection of water at 2.1 microns in the atmosphere of HD 179949 b, as well as the first tentative evidence of night-side emission at 2.3 microns using high-resolution spectroscopy of the same planet. We show that the use of Doppler tomography, in general, greatly reduces background noise levels within the resulting detections, leading the way to its application to fainter signals.

The study of three further planets, HD 189733 b, HD 209458 b and 51 Pegasi b, in the near-infrared enabled the confirmation of previous detections of CO (HD 189733 b, HD 209458 b), H2O (HD 189733 b, 51 Pegasi b) and HCN (HD 189733 b) within their atmospheres. While I was unable to confirm a previous detection of H2O within the atmosphere of HD 209458 b, I am confident that this is not due to the use of Doppler tomography. Instead this may be due to either an error within the atmospheric model used or the potential spurious nature of the previously reported signal.

Finally, the study of the ultra-hot Jupiter WASP-33 b allowed us to extend our use of Doppler tomography to the optical. The pulsating nature of the Delta Scuti host star, WASP-33, had a large affect on the Doppler tomography recovery, leading to a much weaker planetary signal than was expected. However, the cross-correlation recovery was not affected with the same severity. We therefore had to introduce an additional step in the data reduction process to account for and remove the stellar pulsations, which we did by fitting the line profile using least-squares deconvolution. After removing the effect of the pulsations from the observed data, we were able to successfully detect neutral iron within the atmosphere of WASP-33 b.

Throughout this work I have developed processes for data reduction, signal detection and significance calculation when using Doppler tomography. Further work should be focused on improvements to the technique, such as the implementation of a more robust significance calculation, the addition of atmospheric modelling and phase angle modulation. Once this occurs, the use of Doppler tomography will offer an adept way to characterise exoplanetary atmospheres, particularly as we begin to target fainter Earth-like planets.

Date of AwardDec 2023
Original languageEnglish
Awarding Institution
  • Queen's University Belfast
SponsorsNorthern Ireland Department for the Economy
SupervisorChristopher Watson (Supervisor) & Ernst de Mooij (Supervisor)


  • exoplanets
  • exoplanet atmospheres

Cite this