Using starshades to image exoplanets
Over 800 exoplanets—planets that orbit other stars—have been discovered since 1995, and thousands more candidates are currently awaiting confirmation. However, only a small number of very large planets that are far from their parent star have been directly imaged. Direct imaging of exoplanets, particularly Earth-sized bodies in the habitable zone (the region where water can remain liquid, and the temperature and other planetary conditions are conducive to life) of the parent star, is seen as the next major milestone in exoplanet research.1
The diffracted halo of light around a star can be more than 10 orders of magnitude greater than light from orbiting planets (the ratio is known as the contrast), which makes imaging potential exoplanets difficult. Several technologies are being developed that will have the high-contrast detection capabilities required for exoplanet imaging, i.e., that can remove stellar light from a planet's location in an image. It is possible that in the next decade a visible light space telescope with enabling starlight suppression technology for imaging extra-solar Earth-like planets in the habitable zones of nearby stars will be built.
A promising technology uses an external occulter—a satellite that flies far from the space telescope with a large screen, or starshade—to block the incoming starlight. This concept was first proposed in the 1960s and has since been revised in various ways.2–5 However, such starshade designs were not pursued until recently because it was thought that the required partially transmitting screens could not be manufactured. We have recently proposed the use of a flower-shaped screen (see Figure 1) that makes such a mission feasible.6 The screen is derived from the concept of shaped-pupil coronography. By flying a telescope and the starshade we have designed in formation (at a distance of about 40,000km and a holding position to within 1m) at the Sun-Earth L2 Lagrange point, our starshade can create a shadow that suppresses starlight by over 10 orders of magnitude.
A schematic of our starshade, when deployed, is shown in Figure 2. The petals are arrayed around the circumference of a deployable perimeter truss that is similar to those used on large deployable space antennas. The petals themselves consist of a mass-efficient, yet extremely stiff, lattice structure of battens and longerons that intersect a longitudinal spine and a pair of structural edges on each side. These elements are optimized to place and precisely maintain the optical edge with the required profile tolerance, regardless of thermal extremes or structural loads. A schematic of the petal mechanical design is shown in Figure 3. The structural stiffness and petal flatness is maintained by a pair of deployable ribs that fold flat against the petal when stowed, and rise into place when the petal is unfurling. All the system's elements are made from graphite composite material that has a very low coefficient of thermal expansion. The optical edges (25mm wide and 0.4mm thick) have a precisely defined shape that must be maintained to within 20μm and are manufactured separately from the petal. They are made in 10 segments, each between 0.28 and 1.3m long, which is the practical limit for maintaining the required figure profile. The edges are aligned with a coordinate measuring machine and bonded into place on the structural edges.
We manufactured a flight-like, all-composite petal in 2011 and 2012. We began the petal structure assembly with the base spine, then positioned the battens, and finally bonded them to one face of the central spine. We next installed the structural edges, followed by the longerons. Finally, the optical edges, which were divided into five segments for each side, were installed so that they extended about 0.6cm beyond the structural edge. We used a FARO Advantage Platinum portable measurement arm, with an 8ft (2.4m) reach, to position the optical edges onto the mechanical edge. The junction between two edges is shown in Figure 4 and the fully assembled petal is shown in Figure 5.
The final assembled petal shape was measured five times using a coordinate measuring machine. We also conducted a statistical analysis of the measured error on each of the 10 edge segments, over the five measurement sequences. We then performed a Monte Carlo analysis by creating a simulated 30-petal occulter. This simulation included edge segment shape errors for each of the 30 petals, which were derived from the distributions of the 10 measured segments. We used the results of the simulations to construct a probability density of the resulting contrast created by the modeled occulter. The resultant expected contrast value is 2.12×10−11, with a 95% confidence level below 4×10−11. These very low contrasts meet the requirements of space telescope missions for imaging exoplanets.
We have demonstrated, for the first time, the feasibility of building a full-scale occulter to the required tolerances for imaging exoplanets. We are currently working on our second project as part of NASA's Technology Demonstration for Exoplanet Missions program, where we will demonstrate the deployment of a truss outfitted with four subscale petals. We wish to demonstrate that the required 1mm placement accuracy of each petal within the structure is achievable with existing technology. We are also investigating new materials and designs for making sharp optical edges that reduce or eliminate possible glint from the Sun. Our work is one part in a series of technology demonstrations that aim to prove the viability of occulters and could soon lead to an exoplanet detection mission.
This work was funded by NASA grant NNX10AF83G. The work was partially carried out at the Jet Propulsion Laboratory, California Institute of Technology, under contract to NASA.