Revisiting the Borde-Traub focal plane wavefront estimation technique for exoplanet direct imaging

TL;DR

This paper reevaluates the Borde-Traub focal plane wavefront estimation method, demonstrating its theoretical advantages over pair-wise probing, and showcases its practical application in high-contrast exoplanet imaging both in laboratory and on-sky settings.

Contribution

The paper provides a theoretical comparison and experimental validation of the Borde-Traub method, highlighting its efficiency and practical limitations in exoplanet imaging.

Findings

Borde-Traub method can be more efficient than pair-wise probing when convergence time isn't photon-noise limited.

The method has been successfully tested on coronagraphic testbeds at JPL.

First on-sky control of non-common path aberrations achieved with this technique on VLT/SPHERE.

Abstract

Direct imaging of exoplanets relies on complex wavefront sensing and control architectures. In addition to fast adaptive optics systems, most of the future high-contrast imaging instruments will soon be equipped with focal plane wavefront sensing algorithms. These techniques use the science detector to estimate the static and quasi-static aberrations induced by optical manufacturing defects and system thermal variations. Pair-wise probing (PWP) has been the most widely used, especially for space-based application and will be tested at contrast levels of ~1e-9 on-sky along with the future coronagraph instrument onboarding the Roman Space Telescope. This algorithm leans on phase diversities applied on the deformable mirror that are recorded in pairs. A minimum of two pairs of probes are required per bandwidth. An additional unprobed image is also recorded to verify the convergence rate of the correction. Before PWP, Borde & Traub proposed a similar algorithm that takes advantage of the unprobed image in the estimation process to get rid of the pair diversity requirement. In this work, we theoretically show that this latter technique should be more efficient than PWP when the convergence time is not limited by photon noise. We then present its performance and practical limitations on coronagraphic testbeds at JPL and exhibit a first on-sky control of non-common path aberrations with such method on VLT/SPHERE.