PhD defense : Arnaud Striffling

Members of the jury:
Enrico PINNA (INAF, Arcetri, Italy) – Reviewer
Jean-Pierre VERAN (University of Victoria, Canada) – Reviewer
Magali DELEUIL (LAM, AMU) – President
Charlotte BOND (UKATC, Scotland) – Examiner
Simone ESPOSITO (INAF, Arcetri, Italy) – Examiner
Timothy MORRIS (Durham University, UK) – Examiner

Abstract:Direct imaging and the study of exoplanet atmospheres represent a major challenge for ground-based astronomy. The angular separation between the star and its planet is on the order of a fraction of an arcsecond, and the very low flux ratio, ranging from 10⁻⁶ to 10⁻¹⁰, requires the use of high-angular-resolution and high-contrast instruments. Born from a European collaboration, the Extremely Large Telescope (ELT), currently under construction in Chile, should be able to meet this challenge thanks to its resolving power of 3.5 milliarcseconds in the visible domain, provided by its colossal 39-meter-diameter primary mirror. However, as the light wavefront propagates through the atmosphere, it becomes distorted by turbulence, irreversibly degrading the telescope’s imaging quality and resulting in the loss of angular resolution. The addition of an extreme adaptive optics (XAO) system, using a wavefront sensor (WFS) upstream of the scientific instrument, allows for the correction of nearly all turbulence-induced aberrations, restoring optical quality to the diffraction limit. The corrected beam can then feed — notably via single-mode optical fiber — high-contrast or high-spectral-resolution instruments essential for the direct imaging characterization of exoplanets.

The performance achieved by these systems is now so high that it is limited by error budget terms that were previously negligible. This is particularly the case for non-common path aberrations (NCPA), i.e., differential aberrations between the WFS and the imager, which are fully projected onto the latter when the adaptive optics loop is active. While different approaches exist to compensate for these NCPA — such as the use of a static phase plate or an open-loop deformable mirror not seen by the WFS — this thesis explores their introduction through a controlled modification of the WFS operating point. This method is already implemented on the VLT’s SPHERE instrument, which uses a Shack–Hartmann WFS. The extreme correction requirements have led the instrumentation community to turn to WFS technologies known for their very high sensitivity: Fourier-filtering sensors. This increased sensitivity comes at the cost of nonlinear behavior, leading to specific challenges when the WFS operates away from its nominal reference point and to performance losses. To first order, these nonlinearities result in a mode-dependent loss of sensitivity, characterized by optical gains. Knowing these gains is crucial to operating a Fourier-filtering WFS — such as the pyramid sensor — outside of its usual operating point.

This thesis presents an experimental validation, conducted on the PAPYRUS adaptive optics platform, of the until-now theoretical concept of the Gain Scheduling Camera, enabling reliable frame-by-frame estimation of these optical gains. Validating this concept as a robust method made it possible to further explore the possibilities offered by controlled introduction of aberrations. Beyond NCPA correction, mastering an absolute tip-tilt improves optimal injection into a single-mode fiber, while generating dark-hole maps locally increases focal-plane contrast, an essential condition for the direct imaging of exoplanets.