figure [Enlarge image] Source: V. Bailey; WFIRST team / Art by Phil Saunders

Light from the parent star can drown out the signal from orbiting exoplanets.

The Transiting Exoplanet Survey Satellite (TESS) is reaping a huge harvest of “Earth-like” exoplanet candidates (see TESS: The Little Satellite with a Big Job). But directly glimpsing reflected light from a distant planet requires blocking out the light from its parent star, which can swamp the faint exoplanet image. One instrument for doing that blocking is a coronagraph. OPN talked with Vanessa Bailey—an astrophysicist working on the Coronagraph Instrument (CGI) team for NASA’s Wide-Field Infrared Survey Telescope (WFIRST) mission, targeted for a 2025 launch—about how these instruments work.

1. The signal and the noise

The visible light from an Earthlike planet is about 10 billion times fainter than the light from the sunlike star it might be orbiting. Crudely speaking, a coronagraph is like a thumb put in front of the star, blocking its light so that the much fainter planetary light can be analyzed.

WFIRST’s CGI, which will suppress starlight by “only” a factor of a billion, won’t be able to detect true Earthlike planets. Instead, it’s intended to demonstrate technologies that will be needed to do such detection on future missions. Even so, the CGI should enable direct imaging and spectroscopy of Jupiter-like planets and their atmospheres, as well as views of distant, extrasolar Kuiper-belt and asteroid-belt analogs. (Indeed, the CGI will be the first instrument capable of detecting reflected visible light from the cloud tops of Jupiter analogs orbiting other stars.)

2. Adaptive optics

As photons from stars light years away from WFIRST enter the CGI, job one is keeping the telescope pointed at them, and correcting the signal for pointing and “tip–tilt” jitter—to an accuracy of a half-milliarcsecond or less. The first line of defense is a fast-steering mirror, controllable at a kHz frame rate, which passes the tightly aimed light through a focus-control mechanism to a pair of high-actuator-count deformable mirrors. Those mirrors correct for the infinitesimal wavefront errors endemic to the telescope optics.

“The telescope optics are very well polished,” says Bailey, “but they’re nowhere near the performance levels we need—less than a nanometer RMS of residual wavefront errors—to suppress starlight by a factor of a billion.” The DMs and other front-end elements are controlled, at kHz frame rates, by signals from a low-order wavefront sensor that analyzes light from the first “stop” of the coronagraph.

3. Shading the star

The CGI’s business end actually includes three different coronagraph “flavors,” according to Bailey: a “hybrid Lyot” coronagraph (HLC), and two varieties of “shaped-pupil” coronagraphs (SPCs)—bow-tie and disk-shaped. Each channel includes multiple stops to suppress direct and diffracted starlight, and is optimized to work with a specific wavelength of light, so each is paired with a specific color filter at the end of the chain.

The three flavors, Bailey explains, aim at different targets. The HLC, for example, might be used to quickly image and precisely locate multiple exoplanet candidates, with the SPC bow-tie, at a longer-wavelength bandpass, then used for the more time-consuming process of spectroscopy on the most promising candidate. The SPC disk setup, meanwhile, will be useful in imaging distant Kuiper belt and asteroid belt analogs.

4. Final image correction

With the coronagraph stops effectively blocking starlight, an initial “science image” is taken, and the pattern of speckles on it is used to readjust the deformable mirrors, to remove high-order wavefront errors. The result: A clean image of the incredibly faint, reflected light from a planet many light years away, ready for spectroscopic analysis.

5. A packaging challenge

While the picture shown here depicts a linear arrangement, Bailey notes that the WFIRST CGI designers don’t have things nearly so easy—the entire instrument must be packaged into a “not-to-exceed” volume consistent with spacecraft, on an optical bench only around a 1.5 m long. The problem, she says, includes “a lot of different design constraints to weigh” as the team seeks to finalize the instrument’s design—and move to flight hardware development.