Direct observation of Earth-like planets is extremely challenging, because their parent stars are about 1010 times brighter but lie just a fraction of an arcsecond away1. In space, the twinkle of the atmosphere that would smear out the light is gone, but the problems of light scatter and diffraction in telescopes remain. The two proposed solutions—a coronagraph internal to a telescope and nulling interferometry from formation-flying telescopes—both require exceedingly clean wavefront control in the optics2. An attractive variation to the coronagraph is to place an occulting shield outside the telescope, blocking the starlight before it even enters the optical path3. Diffraction and scatter around or through the occulter, however, have limited effective suppression in practically sized missions4,5,6. Here I report an occulter design that would achieve the required suppression and can be built with existing technology. The compact mission architecture of a coronagraph is traded for the inconvenience of two spacecraft, but the daunting optics challenges are replaced with a simple deployable sheet 30 to 50 m in diameter. When such an occulter is flown in formation with a telescope of at least one metre aperture, terrestrial planets could be seen and studied around stars to a distance of ten parsecs.
I thank R. Vanderbei, J. Kasdin, M. Lieber, and J. Arenberg for advice, discussion and encouragement. I thank D. Feldkhun, A. Lo, N. Rajan, E. Schindhelm, and W. Simmons for investigating alternative approaches and verifying the results of this paper. I wish to thank R. Cassanova and the NASA Institute for Advanced Concepts for financial support and encouragement of the work from its early stages.
This file contains Supplementary Information, including Supplementary Figures 1–3 and two references.