Identifying the limits to exoplanets characterization with high-resolution spectro-imaging

Thesis defense of Steven Martos - October, Wednesday 14th at 2.00 pm - Manuel Forestini room IPAG

This thesis is situated within the field of exoplanet detection and characterization through the combination of high-contrast imaging and high-resolution spectroscopy. It focuses more specifically on molecular mapping, a post-processing method based on spectral cross-correlation, designed to extract the signal of faint companions at small angular separations. The overall objective is to identify the physical, instrumental, and methodological limits of this approach, and then to assess its potential on real data and for future instruments.

The first contribution of this manuscript is the establishment of a unified formal framework to describe IFU observations, differential imaging, and molecular mapping. This framework makes explicit the propagation of the useful signal, self-subtraction, the different contributions from fun- damental noise, as well as the role of systematic residuals. It is extended through the development and expansion of semi-analytical tools, FastCurves and FastYield, which make it possible to estimate signal-to-noise ratios, contrast limits, and detection yields as a function of astrophysical, observational, and instrumental parameters. In particular, the manuscript discusses the influence of planetary spectral richness, Doppler configuration, rotation, spectral domain, spectral resolution, spectral format, as well as the fidelity of atmospheric models and the model–reality mismatch.

The second contribution consists in confronting this framework with real data. Using JWST/MIRI MRS observations, the thesis shows that, although molecular mapping makes it possible to overcome a large part of the classical limitation imposed by speckles, real performance is rapidly bounded by other instrumental and extraction-related systematic effects. The analysis highlights the importance of terms such as fringes, straylight, and aliasing, and shows that they can impose a contrast floor in the high-stellar-flux regime. This work is extended by the development of a systematic analysis of MIRI/MRS cubes and by a broader reflection on the exploitation of complex instrumental data, particularly in the contexts of VIPA/PAPYRUS and VLT/HiRISE.

Finally, the thesis opens an instrumental perspective toward future ELT instruments, in particular HARMONI, ANDES, and PCS. It uses FastCurves and FastYield to explore the trade-offs between spectral resolution, spectral domain, sampling, coronagraphy, stability, and systematic residuals, and to translate these trade-offs into performance estimates and scientific yield. One of the key results is that, in the ambitious regimes targeted for the detection of temperate or terrestrial planets, the limiting factor is not only the amount of signal preserved by molecular mapping, but above all the ability of the instruments to deliver sufficiently stable and spectrally clean data. The thesis therefore lies at the interface between exoplanet astrophysics, signal and noise modeling, real-data analysis, and support for the design of current and future instruments.

Thesis direction

• David Mouillet & Alexis Carlotti

The Jury will consist of

• Dimitri Mawet, Full Professor California Institute of Technology, Referee
• Karine Perraut, Astronomer UGA IPAG, Examiner
• Gael Chauvin, Senior researcher OCA, Referee
• Elodie Choquet, Associate Astronomer AMU, Examiner
• Valentin Christiaens, Researcher CEA Saclay, Examiner