GREENBELT, MARYLAND—For months, inside the towering Building 29 here at Goddard Space Flight Center, the four scientific instruments at the heart of the James Webb Space Telescope (JWST, or Webb) have been sealed in what looks like a house-sized pressure cooker. A rhythmic chirp-chirp-chirp sounds as vacuum pumps keep the interior at a spacelike ten-billionth of an atmosphere while helium cools it to –250°C. Inside, the instruments, bolted to the framework that will hold them in space, are bathed in infrared light—focused and diffuse, in laserlike needles and uniform beams—to test their response.
The pressure cooker is an apt metaphor for the whole project. Webb is the biggest, most complex, and most expensive science mission that NASA has ever attempted, and expectations among astronomers and the public are huge. Webb will have 100 times the sensitivity of the Hubble Space Telescope. It will be able to look into the universe’s infancy, when the very first galaxies were forming; study the birth of stars and their planetary systems; and analyze the atmospheres of exoplanets, perhaps even detecting signs of life. “If you put something this powerful into space, who knows what we can find? It’s going to be revolutionary because it’s so powerful,” says Matt Mountain, director of the Association of Universities for Research in Astronomy in Washington, D.C., and former JWST telescope scientist. Like that of Hubble, however, Webb’s construction has been plagued by redesigns, schedule slips, and cost overruns that have strained relationships with contractors, partners in Canada and Europe, and—most crucially—supporters in the U.S. Congress. Other missions had to be slowed or put on ice as Webb consumed available resources. A crisis in 2010 and 2011 almost saw it canceled, although lately the project has largely kept within its schedule and budget, now about $8 billion.
But plenty could go wrong between now and the moment in late 2018 when the telescope begins sending back data from its vantage point 1.5 million kilometers from Earth. It faces the stresses of launch, the intricate unfurling of its mirror and sunshield after it emerges from its chrysalis-like launch fairing, and the possibility of failure in its many cutting-edge technologies. Unlike Hubble, saved by a space shuttle mission that repaired its faulty optics, it is too far from Earth to fix. And not just the future of space-based astronomy, but also NASA’s ability to build complex science missions, depends on its success.
That’s why those instruments sat in Goddard’s pressure cooker for what is known as cryo-vacuum test 3 (CV3). And it is why Webb’s other components—including the mirror and telescope structure, the “bus” that will supply power and control the telescope, and the tennis court–sized, multilayer parasol that will help keep it cool—must undergo a gauntlet of testing, alone and in combinations, until the whole spacecraft is ready. For those on the inside, the strain will only increase as assembly continues, the tests get bigger and more comprehensive, and the spacecraft is launched into space. Only when Webb opens its eye and successfully focuses on its first star will the strain be released.
In the mid-1990s, after Hubble had had its optics corrected and was busy revolutionizing astronomy, researchers began planning its successor. The catch phrase in NASA at the time, championed by agency chief Daniel Goldin, was “faster, better, cheaper.” Goldin challenged NASA engineers and the astronomical community to come up with a follow-on that was cheaper than Hubble but bigger, with a mirror 8 meters across. He received a standing ovation when he described the plans to the American Astronomical Society in 1996. Whereas Hubble covered the whole range of visible light, plus a smidgen of ultraviolet and infrared, the Next Generation Space Telescope (as it was then known) would be a dedicated infrared observatory.
If you put something this powerful into space, who knows what we can find? It’s going to be revolutionary because it’s so powerful.
For astronomers, the infrared spectrum was a beckoning frontier. Visible light from the most distant objects in the universe, the very first stars and galaxies that formed after the big bang, gets stretched so much by the expansion of the universe that it ends up in the infrared range by the time it reaches us. Many chemical signatures in exoplanet atmospheres also show themselves in the infrared region. Yet Earth’s atmosphere blocks most infrared. Webb will give us “the first high-definition view of the midinfrared universe,” says Matt Greenhouse, JWST project scientist for the instrument payload at Goddard.
To capture that light, however, NASA engineers had to overcome huge challenges. The first was heat: To keep the infrared glow of the telescope itself from swamping faint astronomical signals, Webb would need to operate at about –233°C, 40° above absolute zero (40 K). That would require entirely new instrument designs. Size and weight constraints posed additional hurdles: An 8-meter mirror would never fit inside a rocket fairing, so it would have to fold up for launch. The sunshield, too, would have to be collapsible and made of a superthin, lightweight membrane. And the telescope structure would have to be absolutely rigid but lightweight enough to limit the weight of the whole orbiting observatory to no more than 6 tonnes, just a few percent of the weight of a similar-size ground-based telescope. “We knew we would have to invent 10 new technologies” to make the telescope work, says NASA’s JWST Program Director Eric Smith, in Washington, D.C.
Take the mirror. Hubble’s was made from a single slab of glass, but Webb’s folding mirror would need to be segmented, made up of separate hexagonal pieces—a design used in many top ground-based instruments, including the Keck telescopes in Hawaii. The segments would have to be minutely controlled to meld them into a single optical surface, with their reflected light completely in step—a process known as phasing. In Webb, each hexagonal segment will sit on six actuators that control its orientation, plus one in the center to adjust its curvature.
Choosing the mirror material itself was a challenge, because it would have to stand up to a grueling ordeal. Because any material will change shape as it cools, each segment would have to be ground to a shape that is optically wrong at room temperature but warps into one that is correct—to within nanometers—at 40 K. To do that, the mirrormakers planned to combine sophisticated computer modeling with a laborious, iterative process of grinding, cooling, measuring, warming, regrinding, cooling again, and so on. After testing both glass and the metal beryllium, Webb planners chose beryllium because it is strong and light, and it behaves more predictably during repeated cooling and warming cycles.
The final design for Webb fell short of NASA’s original ambitions. Beginning in 2001, concerns about the swelling cost of the telescope forced NASA to shrink the mirror from 8 meters to 6.5 meters, reducing the number of mirror segments from 36 to 18 and its light-collecting area from 50 square meters to 25. But review panels decided that Webb could still achieve its scientific goals. To cut costs further, NASA decided to use less precise mirrors that could be manufactured with many fewer cooling-warming-grinding steps. The change would make Webb less sharp at near-infrared wavelengths between 1 and 2 micrometers—no great loss, as ground-based telescopes already cover that part of the spectrum.
By 2006, all of Webb’s key technologies had been tested and proven viable. The final design was drawn up, and construction of components got underway. Meanwhile, NASA engineers began dreaming up the byzantine series of tests each separate component would have to pass—and the additional tests to be done as components were combined to form larger elements of the spacecraft. “As soon as we put two or three parts together, we test them,” says Scott Willoughby, who is in charge of the Webb effort at Northrop Grumman in Redondo Beach, California.
To put Webb’s enormous mirror through its paces, engineers at the Johnson Space Center in Houston, Texas, completely refitted Chamber A, a huge cryo-vacuum chamber built to test the crew-carrying spacecraft of the Apollo program. For the instruments, they devised the peculiar tortures at Goddard.
The flight models of the instruments began arriving in 2012: four infrared imagers and spectrographs built by collaborators including the European Space Agency, a NASA/European consortium, the University of Arizona, and the Canadian Space Agency. Once the instruments were secured on their rigid framework, they were vigorously shaken to simulate the stresses of launch, as well as blasted with 150 decibels by loudspeaker horns as tall as a person. Next came the first cryo-vacuum test to simulate space conditions.
Problems emerged almost immediately. The heating and cooling caused the delicate multilayer semiconductor sandwiches that make up the infrared detectors to swell and crack. Another critical technology, the microshutter array in the near-infrared spectrograph, also succumbed. This is a device the size of four postage stamps with a grid of 250,000 tiny flaps that can be opened selectively so that the instrument can take separate spectra from, say, 100 galaxies in a single field of view—the first such multiobject spectrograph to fly in space. But the deafening noise of the acoustic chamber caused many of the flaps to jam.
Instrument teams and manufacturers scrambled to identify the problems and produce new parts. Meanwhile, testing went on. All the replacements came together in time for the recent CV3 test, and as the test ended in late January the signs were encouraging that the fixes had worked. “We’re quite pleased with the performance,” says astronomer Marcia Rieke of the University of Arizona’s Steward Observatory in Tucson, principal investigator for the near-infrared camera. “We’re very close to ready for launch.”
While the instruments underwent their ordeal, white-clad engineers in a nearby clean room were painstakingly fitting the mirror segments onto their support, known as the backplane. Hollowed out on the back to reduce weight, each 1.3-meter-wide segment can be carried by a single person, and each has a particular destination on the backplane, depending on its precise optical qualities.
Now that the instruments have been tested and the mirror assembled, these two elements will be mated in March. Then the combined telescope and instrument package, collectively known as OTIS, will endure the shaker tables and acoustic chamber before being inserted into a specially built shipping container. In the dead of night, a truck will carry the container at just 8 kilometers per hour from Goddard to Joint Base Andrews, where it will be placed into a huge C-5 Galaxy transport plane, with just centimeters of clearance, for its flight to Houston.
The few months OTIS spends in Chamber A early next year will be the most critical it will face. Light sources on the ceiling will create an artificial universe, allowing NASA engineers to run light all the way through the system from main mirror to detectors for the first and only time in spacelike conditions. They will practice phasing up the mirror and will check out all observing modes of the four instruments. “Hubble didn’t do an end-to-end optical test. We’re not skipping that on this program,” Greenhouse says.
Then it’s back into the shipping container and another C-5 flight to Redondo Beach, where Northrop Grumman has been building the bus and sunshield. There the full observatory will take shape as the telescope and instruments are mated to these last two elements.
Now too large to fit inside a plane, Webb will make its final prelaunch journey by ship, down the California coast and through the Panama Canal to French Guiana—home of Europe’s spaceport, and a waiting Ariane 5 launcher, part of Europe’s contribution to the project. In October 2018, the Ariane will fling Webb toward L2, a gravitational balance point 1.5 million kilometers from Earth, directly away from the sun. The journey will take 29 days.
Webb will begin unfolding and deploying components almost as soon as it hits space. Deployment will be “3 weeks of terror,” Mountain says. “No one has done this before, ever.” First to deploy will be solar arrays and antennas to provide power and communications with Earth; then the sunshield will unfurl to begin cooling the telescope and instruments; finally, the secondary mirror will swing into position and the main mirror wings will snap into place. Once the mechanical gymnastics routine is finished, there will come the heart-stopping moment when the mirror first looks at the sky. Then the mirror has to be phased up, and the instruments cooled and all their modes tested. Commissioning is expected to take a full 6 months after launch.
“A whole chain of things have to be done to get that really good-looking star,” says Lee Feinberg, JWST telescope manager at Goddard. “But then we can really rest.”
Until then, the pressure will be unrelenting. But the builders of Webb say they do find time to reflect on what they are doing. Pierre Ferruit, JWST project scientist at the European Space Agency in Noordwijk, the Netherlands, recalls watching from the control room at Goddard during CV3 as technicians carried mirror segments into the clean room and fitted them to the backplane. “Even for someone working on the mission, it’s quite incredible,” he says. Rieke had the same sensation: “It’s just enchanting to be witnessing history.”