Experimental Demonstration of Wavefront Estimation in a Shaped-Pupil Coronagraph

Direct imaging of extrasolar planets, and terrestrial planets in particular, is an exciting but difficult problem requiring a telescope imaging system with unprecedented levels of contrast. One promising design is the Shaped Pupil Coronagraph (SPC), pioneered by our lab over the past several years. The SPC was designed to achieve $10^{10}$ contrast at an inner working angle of 4 $\lambda/D$, based on the requirements of NASA’s space-based Terrestrial Planet Finder Coronagraph (TPF-C) mission. However, it has long been recognized that a key problem in achieving these requirements in practice is estimation and control of wavefront aberrations in the optics of the telescope. Furthermore, it is crucial to correct as fast as possible because of finite mission lifetime as well as due to the non-static nature of some aberrations. In earlier work, we have used the so-called speckle nulling algorithm to achieve $10^6$ contrast at 4 $\lambda/D$ in air on the Princeton testbed, as well as almost $10^8$ contrast in vacuum on JPL's testbed. However, the speckle nulling algorithm does not fully estimate the wavefront, and is therefore quite slow, requiring hundreds or even thousands of iterations. In this work, we present experimental results of wavefront correction using wavefront-estimation based techniques, which lead to much better performance and speed. In particular, we use a variant of the "peak-a-boo" algorithm as well as the Borde-Traub algorithm, both of which work with image-plane data and use different patterns on the DM to introduce diversity. Finally, we study the performance at different wavelengths and in broadband light.