Skip to main content


 June 4, 2019

James Webb Space Telescope will rely on Fienup’s algorithms to see clearly

A group photo of Tang, Fienup, and Paine.
From left to right: PhD candidate Joseph Tang, Professor James Fienup, and PhD candidate Scott Paine in the lab where they use a MEMS (micro-electro-mechanical systems) deformable mirror to simulate the James Webb Space Telescope and test the phase retrieval algorithms the lab has developed to realign the telescope’s 18 mirrors.

A memorable scene from the movie Apollo 13 shows three astronauts being relentlessly bounced and jostled during their launch. Imagine what all that bouncing does to a finely tuned space telescope when it is lifted off a launch pad!

Especially if the telescope is as complicated as the James Webb Space Telescope (JWST). It will be packaged like a table with folding ends for its launch two years from now, then unfolded once it is in orbit—at which time 18 individual hexagonal shaped mirrors will need to be carefully aligned to nanoscale tolerances to deliver sharp images.

NASA will use phase retrieval algorithms developed by James Fienup, the Robert E. Hopkins Professor of Optics at the Institute of Optics, to do that alignment. The algorithms compare the blurry images of a bright reference star taken by the newly deployed telescope to how the star should actually appear—and adjust seven actuators on each of the mirrors accordingly.

And just in case, NASA is funding Fienup’s group to develop even more robust algorithms for worst-case scenarios: If JWST’s mirrors need to be subsequently realigned on short notice, for example, or at more frequent intervals than expected.

“We’re on what they call a risk reduction effort,” says Fienup. “They want to use what we’re developing in case they have to correct the alignment every two hours. You couldn’t take the time to do a big telescope slew to look for a particular bright star. You’d have to use whatever happens to be in the field of view of the telescope at the time. It might be a dim star, or maybe a galaxy. That’s much more complicated.”

James Fienup standing in front of the James Webb Space Telescope in 2017, just before it left Goddard Space Flight Center for testing at Johnson Space Center. 
James Fienup standing in front of the James Webb Space Telescope in 2017, just before it left Goddard Space Flight Center for testing at Johnson Space Center.

Fienup is one of the world’s leading experts in phase retrieval. His paper, “Phase Retrieval Algorithms: A Comparison,” is the most highly cited paper (out of over 50,000) in the journal Applied Optics, according to Web of Science.

He first demonstrated the value of phase retrieval algorithms for space telescope aberration correction when he contributed to an effort that corrected the Hubble Space Telescope’s infamous “nearsightedness”  (spherical aberration) when it was launched in 1990. At the time he was a scientist at the Environmental Research Institute of Michigan.

Starting a year after joining the University of Rochester in 2002, he began receiving a continuous stream of grants from NASA to refine the algorithms not only for the JWST, but for future telescopes including NASA's Wide-Field Infrared Survey Telescope (WFIRST). WFIRST is not segmented. However, precise knowledge of its small wavefront aberrations, which can be determined with phase retrieval, will be required for exoplanet direct imaging and microlensing. And exquisite wavefront control and highly accurate modeling will be required for NASA's proposed Large UV/Optical/Infrared (LUVOIR) Surveyor telescope, which is segmented and will be able to image dim “exoplanets” orbiting bright stars.

“The vast majority of our work is math, computer algorithms, and computer simulations,” Fienup says, “but we also have a laboratory where we use a device that simulates the JWST—a little MEMS (micro-electro-mechanical systems) deformable mirror that has a bunch of little hexagons that can move in and out and tip and tilt. So, we have laboratory experiments going on also. That gives us a step in between pure simulations and doing it in space.”

Two of his PhD students—Matthew Bergkoetter ’17 and Scott Paine—traveled to the Johnson Space Center in 2017 to complete analysis of data gathered during cryogenic testing of the JWST. The students spent four weeks in Houston, weathering the rains of Hurricane Harvey as they assisted NASA personnel in measuring residual wavefront error in the fully-assembled telescope. JWST was stored in a cryo-vacuum chamber that simulates the environmental conditions of space.

The students employed image-based wavefront sensing techniques to measure the observatory’s pupil phase, which was used to verify that the as-built alignment of the optical elements meets specifications. The testing of the various science instruments was done a couple of years earlier at NASA's Goddard Space Flight Center, assisted by former PhD student Dustin Moore '16. Another student in the group, Joseph Tang, will hopefully help with the commissioning of JWST after launch.

Bergkoetter, along with three other PhD students from the Fienup group, is now an optical scientist at the Goddard Space Flight Center.

Fienup’s group is engaged in other aspects of unconventional imaging—“where you’re trying to form images, and you have to do something unusual to make it work,” he says.

For example, his group uses imaging interferometry to capture fine resolution images of satellites orbiting the earth 37,000 kilometers overhead—from a collection of ground-based telescopes. Applications include being able to assess whether a “dead” satellite has salvageable parts that could be transferred to a newly launched satellite, cutting costs.

His group also is working on ways to overcome the blurring that occurs when attempting to capture images horizontally, through the atmosphere near ground level, where turbulence is most severe. Another current project is to modify the approaches developed for space telescopes so they can be used for measuring freeform optical surfaces during their manufacture.