Ultra high precision wavefront sensing for Extreme-AO on ELTs

Abstract

High precision wavefront sensing is the key to detect low mass (potentially rocky) planets with ELTs. Ideally, a wavefront sensor for an Extreme-AO system should be both very sensitive (to allow high speed wavefront correction) and very accurate (to allow precise calibration of residual starlight vs. planet light in the focal plane). I describe two options which meet these requirements: (1) non-linear curvature wavefront sensing is several orders of magnitude more sensitive than conventional WFSs, and can work at full sensitivity in open loop or in the visible (2) focal plane wavefront sensing combines high sensitivity and is free from non-common path errors. It can also measure light coherence, and theref ore separate speckles from planets. Combining these two schemes is especially attractive for Extreme-AO systems aimed at direct imaging of exoplanets with ELTs. Laboratory demonstration of both Focal plane wavefront sensing is also be presented. 1 Wavefront sensor sensitivity The Wavefront sensor sensitivity is defined as its ability to use a limited number of photon for accurate measurement of wavefront aberrations. In an ideal WFS (optimal sensitivity), for each wavefront mode to be measured, the error in radian RMS on pupil is equal to the inverse square root of the number of photon. This simple equation can be generalized to a wavefront measurements of M modes with a total of N photon: (rad) = p M/N (1) Direct comparison between this ideal limit and current wavefront sensing schemes reveals a huge performance gap, especially at low spatial frequencies. Commonly used wavefront sensors such as Shack-Hartmann and Curvature are very robust and flexible, b ut are poorly suited to high sensitivity wavefront measurement: for low-order aberrations which prevent high contrast imaging very close to a star they require ((p/2)/r0) 2 (with p = spatial period of the aberration) more photons than the theo- retical limit above. This is a factor 100 to 1000 on 8 to 10-m telescopes. These conventional wavefront sensing schemes o er nearly ideal sensitivity at a single spatial frequency (d efined by the WFS sub- aperture spacing for SH and curvature) but su er from poor sensitivity at low spatial frequencies (an e ect commonly referred to as noise propagation), which are most critical for direct high contrast imaging of exoplanets and disks. WFS schemes o ering very high sensitivity across a wide range of spatial fr equencies have previ- ously been proposed: Mach-Zehnder pupil plane interferometry (Angel 1994), Zernike phase contrast wavefront sensing (Zernike 1934, Bloemhof & Wallace 2003), pyramid wavefront sensing (Ragazzoni 1996). These WFSs concepts do indeed o er sensitivity at or very close to the theoretical ideal limi t across a wide range of spatial frequencies (Guyon 2005), but only in the small aberration regime where the wavefront error at the sensing wavelength is below 1 rad RMS (= 87 nm RMS in V band) at the sensing wavelength. In§2, I describe a new type of wavefront sensor which, for the firs t time, o ers full sensitivity with- out requiring the wavefront to be nearly flat at the sensing wa velength. The approach is derived from the already highly successful curvature wavefront sensing technique, but uses a non-linear wavefront reconstruction algorithm (non-linearity is unavoidable f or simultaneous delivery of high sensitivity