At NASA’s Goddard Space Flight Center in Greenbelt, Maryland, the agency’s biggest-ever science project is coming together at last.

In a gymnasium-sized cleanroom dominated by a laser-guided robotic arm on mustard-yellow scaffolding, bunny-suited technicians have completed the primary mirror of the James Webb Space Telescope, a $9 billion orbital observatory planned to lift off in 2018, being built by NASA in partnership with the European and Canadian Space Agencies. Gripped by the robotic arm and guided by technicians, the last of the mirror’s 18 lightweight hexagons of gold-coated beryllium has been installed, marking the most tangible milestone yet in the observatory’s multi-decadal path to launch.

Each segment is as big as a coffee table but hollowed out to only weigh 20 kilograms. The entire mirror spans 6.5 meters edge to edge. After being mated to the rest of the telescope, which is still under construction, it will ultimately be launched to its deep-space destination some 1.5 million kilometers from Earth. There, at a point of gravitational quiescence called L2, Webb will begin what astronomers say will be revolutionary studies of the universe.

Turned skyward and concave in a supportive cobweb of carbon fiber called a backplane, the mirror looks like the giant, unblinking compound eye of an insect. For Alan Dressler, a senior astronomer at the Carnegie Institute for Science, Webb’s completed mirror brings other anatomies to mind. “The mirror is the heart of a telescope,” he says. “To finish Webb’s mirror is to hear the first heartbeat of the magnificent creature that will carry us to when the universe that bore us was itself born.”

That may sound overly dramatic, but, then again, the universe is a dramatic place, and Dressler has been waiting for this moment for more than twenty years. In the early 1990s, after the launch of NASA’s Hubble Space Telescope, he chaired an influential committee that recommended the agency’s next great observatory should do something Hubble could not—stare back all the way to nearly the beginning of time, to “first light,” an era more than 13 billion years ago when stars and galaxies first coalesced and ended the primordial dark ages of the cosmos. Witnessing that first light, astronomers could then retrace the universe’s evolution in unprecedented detail, watching the assembly and growth of galaxies, the emergence of subsequent generations of stars and the births of planetary systems. Now, finally, the Webb telescope is poised to do just that.

To see all the way back to the first lights turning on in the universe, the telescope needs a mirror much bigger than Hubble’s 2.4-meter disk of silvered glass. It must also be very cold. As it streams through the expanding universe, the visible light emitted by those very first luminous objects is stretched like taffy, becoming a ghostly infrared glow that can only be seen—felt, really—by something cooled close to absolute zero, the coldest temperature there is. To conceive of how difficult it is to glimpse these distant galaxies, imagine looking up at the moon and trying to find the faint glow of a child’s nightlight on its surface. In its search for first light, Webb’s planners say, it will be trying to see galaxies twenty times dimmer than that.

As Webb edges closer to ushering in a new era of space science, one might wonder where it came from, what exactly it will do, and what, if anything, might come after its mission is done. As the last segments of the observatory’s giant mirror were being set into place, I visited Goddard to talk with the people who know Webb best—the scientists and engineers who have brought it to life through a long, rocky and multigenerational gestation.

A Tangled Webb
Webb’s foremost scientific champion is arguably John Mather, a slender and soft-spoken astrophysicist at Goddard who wears sensible shoes and a toothy grin. He speaks about Webb like a grandfatherly Scoutmaster would about building fires or tying knots—there is a soothing, almost wholesome patience in his diction, and he savors cutting through difficult details with short, simple summaries. It also does not hurt that he has a Nobel Prize in physics, which he won for groundbreaking studies of the cosmic microwave background, the faint afterglow of the Big Bang that constitutes the universe’s first and earliest baby picture.

Mather was still in his forties when he became Webb’s senior project scientist in 1995, and today is nearing 70. In all that time, he has been working toward placing the next set of pictures in his cosmic photo album, the telescope’s promised images from the mysterious era of first light. No one really knows what the first stars looked like, or whether the universe’s first luminous objects were even stars at all. Instead, Mather says, the first light could have come from supermassive black holes that were messy eaters. Found at the centers of most galaxies, these billion-solar-mass behemoths must have plumped up by swallowing immense volumes of gas in the primordial universe, building up white-hot accretion disks around their maws like barbeque sauce on the jowls of a competitive eater. The universe’s first light could have come via glowing crumbs from a black hole’s table, and if so, Webb could tell us.

These days, however, Mather is most excited about what Webb can reveal of our local, present-day corner of the cosmos, rather than its far-off past. The observatory’s keen infrared vision can peer into the dust-shrouded centers of molecular clouds and circumstellar disks to watch as worlds grow like embryos in a womb. And Webb’s giant mirror, he says, is just big enough to spy signs of water vapor—evidence of possible oceans—upon a few favorably positioned small, just-maybe rocky and Earth-like planets around our nearest neighboring stars. “Ever since I was six or seven years old, I’ve been wondering ‘how did we get here?’ but I couldn’t get the answers,” Mather says. “We didn’t know anything about the primordial universe. We didn’t know if planets were unique to our sun or common. We still don’t know if life is unique or common. Every step along that path is something Webb can work on.”

Yet despite all the transformative science Webb promises, its development has been plagued by difficulties. Initially projected to cost less than a few billion dollars and to launch as early as 2011, repeated schedule slips and budget overruns soon made Webb one of NASA’s most troubled and troubling programs. As other, smaller NASA astrophysics projects suffered postponements and cancellations to offset Webb’s swelling costs and delays, the observatory earned a reputation as “the telescope that ate astronomy.” Astronomers wondered whether Webb, named for a NASA administrator who led the agency through its Apollo heyday, would ever actually fly, and, if it did, whether its ballooning cost would make it the last gasp for the agency’s program of great observatories.

The troubles culminated in 2010 and 2011, when fed-up members of Congress threatened to defund Webb entirely. Thanks to an independent review followed by a program-wide “replan,” Webb received more money and survived. In the few years following this reform, Webb has stayed within its new budget and on target for a launch in October 2018. “We had a near-death experience,” Mather says. “We had to tell Congress the budget wasn’t enough, and they were either going to kill Webb or make it right. From my perspective, a miracle happened.”

Part of the trouble early on was that Webb was relying on wholly new technologies—such as its large cryogenic mirror— that were destined to take longer than most people expected. “We came to recognize fairly quickly that [making Webb] could take about 20 years,” says Garth Illingworth, an astrophysicist at the University of California, Santa Cruz who has been heavily involved in planning the telescope. “But 30 years—that was a bit long for those of us in our early forties!”

Making the Mirrors
Compared to old salts like Mather and Illingworth, Lee Feinberg, an engineer at Goddard who oversees all of Webb’s optics for NASA, is a relative newcomer to the project—a fact best evinced perhaps by his full head of dark bushy hair without a touch of gray. He arrived in 2001, when his daughter was a toddler. Now, she is about to graduate high school, and will be nearly out of college when the telescope launches. “After all these years, my kids kind of feel like the telescope is their parent, too, that our family includes some guy named Webb!” he jokes.

Part of Feinberg’s job was figuring out how to make Webb’s 6.5-meter mirror suitable for launching into space. To fit inside a rocket, the giant mirror had to be segmented and stowable, so that it could be folded and unfolded like a piece of origami. And though it would be almost three times larger than Hubble’s mirror, with nearly seven times as much collecting area, it had to weigh much less. Glass is relatively easy to work with, but is also heavy and not very resilient to cryogenic temperatures, so Feinberg and his colleagues chose beryllium mirrors instead—one of several innovations that make the entire observatory a featherweight, with less than half of Hubble’s total mass. In some cases, Webb’s builders had to develop new technology just to confirm other new technologies worked, like the cold-resistant optical systems for monitoring mirrors inside cryogenic tanks, or the laser metrology platforms that measure and guide the precise sculpting of mirror surfaces during polishing.

In the end, so much effort was put into the mirrors that they came out ahead of schedule. After a production process that took them zigzagging across the country between specialized laboratories scattered through eight different states, the mirrors have been sitting in storage at Goddard for the past two years, waiting for the completion of other lagging components necessary for their assembly. Finally, though, all of the elements for Webb’s heart have come together. “When I started working on this, I was told it was going to be a marathon, not a sprint, but I feel we’ve just sprinted a marathon,” Feinberg says. “The assembly of the primary mirror is a huge milestone that really shows we’ve conquered every possible obstacle we’ve encountered so far.”

Even so, Webb has a long way to go before reaching the launch pad and L2. No one has ever built a space telescope so big before, let alone one meant to deploy and operate at such cold temperatures so far from Earth. Today, the greatest cause of worry is probably Webb’s sunshield, a plastic parasol made of five tennis-court-sized layers of ultrathin Kapton film designed to block the sun and cool the telescope down to its operating temperature of 50 degrees Kelvin—about the average surface temperature of Pluto. Like the observatory’s mirrors, the sunshield will also be stowed for launch, then remotely released in space. The sunshield deployment will be a figurative and literal high-tension process in which a fiendishly complex system of actuators, pulleys and wires holds each layer taut and flat as it unfurls, all to eliminate any kinks or tears in the material that could ruin the mission. Many Webb veterans compare the sunshield’s deployment to jumping out of a plane with a parachute strapped to your back that someone else has packed—how can you really know it will open?

That “someone else,” in this case, is the aerospace company Northrop Grumman, NASA’s primary contractor for Webb. Across the country from Goddard, in Redondo Beach, California, the company’s technicians are assembling and testing the sunshield—paying particularly close attention to its packing and deployment. Eventually, all the telescope’s parts will be taken via truck, rail and air to be mated together in Redondo Beach. Webb will then be loaded aboard a barge that takes it through the Panama Canal and to the northeastern coast of South America, where it will launch from a spaceport in French Guiana on an Ariane 5 rocket provided by the European Space Agency.

As its finished hardware at last comes together for the final push to launch, almost everything about the telescope and its construction appears majestically, almost comically outsized. Except, that is, its margin for error. Hubble was the last time NASA attempted so many great technological leaps in one project, and it was almost dead-on arrival after its primary mirror turned out to be incorrectly polished. Astronauts rode space shuttles up to Hubble to repair it, and subsequent servicing missions repeatedly upgraded the telescope’s failing hardware. But when Webb launches, there will be no Plan B—most of the telescope’s costs come from tests its builders must run to prove all of its new technologies will work as intended.

Will It Work?
“You want to know why Webb costs so much?” asks Mike Menzel as we gaze at the mirror from a viewing area overlooking the cleanroom. “I’ll show you.”

Menzel, NASA’s systems engineer for Webb, has been working on the telescope since 1997. If you didn’t know he was an engineer, you could guess by his precisely combed gray hair, eternally arched eyebrows and the technical wisecracks he utters from behind a dense, well-manicured beard. Other than Webb’s overall project manager, Goddard’s Bill Ochs, Menzel is the NASA official responsible for every aspect of the telescope’s design and testing, right down to each individual bolt, circuit, and cable across the entire observatory.

He ushers us away from the cleanroom, through a series of hallways and doors. We pass by a swimming pool-sized cylinder nestled in coiled piping and vapor-billowing valves. It’s a cryogenic vacuum chamber that simulates deep-space conditions, and the telescope’s science instruments are undergoing a lengthy test inside. We pass other rooms lined with vibrating “shake tables” and gigantic air horns, chambers where the assembled mirror, the instruments, and other telescope components will be subjected to jostling G-forces and acoustic shockwaves that mimic a rocket ride to orbit.

We finally reach Menzel’s office, where he opens a file on his desktop computer and displays it on a nearby flatscreen. The screen shows a beryllium mirror segment laying motionless upon a shake table, embedded with hairsbreadth tolerances within a tangled nest of carbon-composite actuators and electrical wiring from the telescope’s backplane. Mounted to the back of a segment, the actuators will serve to adjust the segment’s position to nanometer precision once in space.

Beryllium was chosen for the mirrors not only because it is lightweight, Menzel says, but also because it scarcely warps at cryogenic temperatures. Still, warping cannot be completely eliminated, so each segment is the product of a painstaking polishing process to precisely counteract the warps—technicians freeze the segments, measure the warps, bring them back to room temperature, then polish the surface to nullify them. If at “cryo” a portion of a segment’s surface forms a nanoscale hill, at room temperature the hill is then polished into a valley of exactly the same dimensions. When chilled close to absolute zero again, the valley will warp into a plain, and the mirror segment will be flat. The process yields an average surface error across the coffee table-sized segment of only 25 nanometers—one four-thousandth the thickness of a single sheet of notebook paper. Scaled up to the size of the continental United States, a cryo-frozen mirror segment’s biggest topographical feature would be just less than 8 centimeters high.

“So this thing takes almost five years to build, and it’s stable to 25 nanometers,” Menzel says. “Then I give it to our structural analyst and here’s what she does with it.” The image on Menzel’s flatscreen begins to move. The shake table activates, and the mirror segment begins to violently oscillate back and forth, until it looks like a motion-blurred slice of jiggling pineapple Jell-O. I notice I’m clenching my teeth, as if they’re going to rattle out of my skull just by watching.

Menzel smiles. “If she didn’t do that, the launch vehicle would. Some of those components are feeling 11 G’s here, but when it’s all over, they and the mirror are just as good as they were before. Going into space is just hell on wheels. And unlike Hubble, which we built stable as a brick outhouse and launched into orbit as-is, here we build a beautiful telescope and prove it’s great on the ground, then we collapse it all down, launch it and reassemble it in deep space. That’s why this is expensive.”

The launch isn’t what worries Menzel, though—he sweats the deployment. He has walked through the sequence in his head countless times. A half-hour after launch and 10,000 kilometers away from Earth, the telescope separates from its booster. It deploys its solar-power arrays and communications antenna en route to the Moon, which it passes two and a half days after launch. Over the next two weeks, Webb slowly deploys its sunshield and mirrors in a series of motions so delicately choreographed they seem worthy of a symphony.

“As soon as we pass the Moon, my blood pressure spikes,” Menzel says. “We all joke around about the Curiosity rover going to Mars and the ‘7 minutes of terror’ its team talked about for atmospheric entry, descent and landing. Well, we have two weeks of terror as we watch things deploy. If it all is well after two weeks, that’s where I get lost and go on a bender.”

Countdown to Launch
For Menzel, Webb reaching first light will fulfill a childhood dream, and make the long years of stress worthwhile. “When I was a kid, I wanted to be an astronomer and build my own telescopes, and the world’s biggest telescope [at the time] was near where I grew up, the Hale Telescope on Mount Palomar. It has a mirror 5 meters across, and it weighs half a million kilograms. It’s a monster. Well, Webb’s mirror is bigger than that, and it only weighs 6,620 kilograms. Excuse me, 6,338—it’s actually underweight! And we’re sending it beyond the moon, to see the very first things that turned on in the universe.”

Webb’s mission is only slated to last five years, but Menzel and Mather are confident that by carefully managing the observatory’s onboard reservoirs of propellant, which are used to stabilize its position at L2, they can operate Webb for at least ten. The gradual degradation of instruments and other hardware under the incessant bombardment of micrometeorites and cosmic radiation is more likely to end the mission than anything else, Menzel says. Based in part on lobbying from Frank Cepollina, a senior engineer at Goddard who pioneered the in-orbit servicing of satellites, NASA added features to Webb's docking ring to accommodate a rendezvous with astronauts, or, more likely, a robot, but such missions at present seem too risky, expensive and premature to merit serious consideration.

With or without a servicing mission, if Webb manages to survive for more than a decade, eventually its orbit at L2 will briefly, fatefully align with Earth’s shadow, cutting off solar power to the observatory for a few hours—long enough for its onboard batteries to drain and potentially die.

NASA is already working on what comes next after Webb, a more modest Hubble-sized observatory called WFIRST-AFTA (don’t ask) designed to study dark energy and other planetary systems. If all goes according to plan, WFIRST could launch as early as 2024 for as little as a quarter of Webb’s high cost. Beyond that, astronomers are already drawing up plans for an 8- to 16-meter segmented broadband telescope after WFIRST—as fantastical as it seems, Mather, Menzel and other experts believe leveraging Webb’s now-mature technology could make building that behemoth no more expensive than Webb itself. Such a telescope could look for twins of Earth around hundreds or thousands of nearby stars, seeking out worlds graced with oceans, clouds, continents, and just maybe, beings staring back at us through space telescopes of their own.

In this view, gazing at Webb’s completed primary mirror awaiting integration and launch in the sterile confines of a cleanroom is really to glimpse the beginning of the future, the first tenuous glimmer of light on the horizon before the dawn. “Webb’s mirror is built on the shoulders of giants,” says Matt Mountain, Webb’s telescope scientist and president of the Association of Universities for Research in Astronomy. “Assembling the mirror is the culmination of four centuries of progress that began with the telescopes of Galileo and Newton. We are now about to fly the largest space observatory humanity has ever built. For those of us who make telescopes, the only thing more humbling and exciting is imagining what comes next.”