How NASA Scrambled to Save OSIRIS-REx From Leaky Disaster
ON OCTOBER 20, an uncrewed spacecraft roughly the size of a Sprinter van and traveling at the glacial pace of 10 centimeters per second collided with an asteroid 200 million miles from Earth.
The OSIRIS-REx craft’s proboscis-like “Touch-and-Go Sample Acquisition Mechanism” (Tagsam), an 11-foot-long shock absorber tipped by a round vacuum head and a collection canister, touched down atop a boulder on the asteroid Bennu’s surface—and appeared to smash right through it. Several seconds after impact, the arm had punched more than a foot and a half into the asteroid. It would have kept going too, but for the programmed sequence that burst the arm’s nitrogen gas canister, tripped its vacuum suction, and milliseconds later fired the spacecraft’s reverse thrusters to initiate a hyperbolic escape trajectory. After 17 years and $800 million in funding, the crux of OSIRIS-REx’s smash-and-grab mission was over in 15 seconds.
At the Lockheed Martin mission support area in Colorado, the masked OSIRIS-REx team celebrated with air hugs and elbow bumps. A clear landing spot, flush contact, and deep penetration into Bennu looked good to ops. The Tagsam should have collected more than its target mass of 60 grams (about 2 ounces, or one large egg’s worth) of fragments, dust, and rocks (“regolith,” in asteroid-speak) from the surface, perhaps much more. All that remained was to perform a sample mass measurement maneuver to estimate the collection’s actual mass, stow the collector head away in a sample return capsule, and navigate it back to Earth by the year 2023. Then, scientists could begin to study the space dust to learn more about the beginnings of the universe, whether asteroids might have brought water or even life to Earth, and how to react if Bennu turns out to be on course to hit Earth between the years 2175 and 2195, as it seems it might be. In the middle of a global pandemic, the touch-and-go felt like a triumph.
But several team members had nagging concerns. “There was almost no resistance at the surface,” says Mike Moreau, OSIRIS-REx’s deputy project manager. “When the gas bottles fired, it looked like it blew all the surface material away like it was packing peanuts.”
Two days later, the team repositioned the Tagsam arm in front of a ship-mounted camera for visual inspection. The head appeared to be packed full of material—but it also appeared to be leaking. A series of three still images projected in sequence onto a large screen at Lockheed MSA showed a cloud of rocky material escaping into space. The room began to buzz with nervous discussion. The next photos, long-exposure images, appeared to show the precious asteroid debris leaving the Tagsam head’s one-way Mylar flap like water streaming from a showerhead. Dante Lauretta, the mission’s principal investigator, shouted over the chatter: “We have to do something about this!”VIDEO: NASA/GODDARD/UNIVERSITY OF ARIZONA
But the delay in sending and receiving images from 200 million miles away meant the leak had actually happened 30 to 40 minutes before. The team huddled in groups and began to ask questions about just what the hell had happened. How long had the arm been leaking? How much material had escaped, and why? How could they stop it?
The leak’s source was easy to spot. Rock-sized pieces of regolith were bulging the head’s Mylar flap ring partially open in several places. The flap was meant to allow material in—but not out. Nothing like this had happened during testing, which had included simulations of near-zero-G conditions using regolith-like materials, says Beau Bierhaus, Lockheed Martin’s Tagsam lead scientist. The particles appearing to hold the flap open were the right size and shape for collection. “I can’t think of anything that would have prevented the particles from being collected [inside the Tagsam head], other than there was no more room left at the inn,” Bierhaus says. “Because there was no more room inside, it got stuck.”
How might the Tagsam head have become so full? Because Bennu’s surface was a mystery to scientists before OSIRIS-REx arrived to scope it up close, Bierhaus and other Lockheed engineers had to design their collector head to bounce off and suction up a range of surface types, from ones similar to a hard-packed gravel driveway to ones softer than a fine, sandy beach. Before the team saw Bennu up close, they modeled its surface based on the 25103 Itokawa asteroid, sampled in 2005 by the first Japanese Hayabusa mission. “We were hoping to, in essence, scoop up a big bucket of soft sand,” says Ed Beshore, the former deputy principal investigator of the mission, now retired from the University of Arizona. Instead, pictures of Bennu’s surface taken by OSIRIS-REx’s cameras before the touch-and-go appeared to show a minefield of sharp boulders and rocks.
But Bennu had more surprises in store. In fact, based on the Tagsam’s deep bounce, it seems the surface material was not hard. In the asteroid’s microgravity environment, it instead behaved like a viscous fluid—thousands of marbles bouncing and scattering in low gravity. “If you push into it, it displaces and moves in ways we could not have anticipated,” Bierhaus says.
The head penetrated the first few centimeters of surface without much resistance. This, Moreau says, “preloaded the center of the Tagsam head with material, and then when the gas blew, all that stuff went into the head immediately.” As the arm continued downward another half meter through the yielding surface, more regolith might have been jammed in. “By the time we backed away, the head would’ve been packed full,” he continues. Another possibility, given the surprisingly viscous surface material, is that the regolith’s soft, malleable rocks wedged into the Mylar flap opening and weren’t able to make it all the way into the head, Moreau says.
Still, at HQ, there was some good news. Twenty to 30 minutes after the spacecraft stopped moving its Tagsam arm, the leak of material appeared to have died down. “Every time we moved the arm, we were shaking stuff loose,” Moreau says. Now the team ordered the ship to quiet itself, point toward Earth for easy communication, and “park” its arm in place. The team also canceled the upcoming sample mass measurement maneuver, which required extending the Tagsam arm and spinning the spacecraft—an action that was likely to spray debris out of the head in 360 degrees.
Confident the Tagsam had gagged up only a portion of its enormous bite, the team moved on to the next question: Assuming the head had been crammed full of material when it bounced off Bennu, and that the leakage had been caused largely by movement of the arm, how much of the sample had been lost? Were there at least 60 grams left to stow away?
To answer those questions without the measurement maneuver, five teams set about making estimates using alternate techniques. One group analyzed high-resolution imagery of the landing zone, down to the individual rocks, to model how many grams should have been collected; they estimated it was likely hundreds. Another group pored over photos of the Tagsam after the touch and go, peering into its visible area (about 40 percent of the container) to estimate the volume of the debris inside. The obstruction of light seen in a screen ringing the outside of the container offered another clue that the capsule might be close to full. One team estimated that the rocky material jammed in the Mylar flap was in the tens of grams—not enough to make up the necessary sample on its own, but a sizable prize. Another team used new 3D imaging techniques to estimate the size and mass of hundreds of particles shown escaping during the 10-minute imaging session just after the movement of the Tagsam arm, and found loss in the tens of grams—a “decent amount,” says Coralie Adam, the mission’s lead optical navigation engineer, but “we probably lost the smallest material that could escape through those gaps.”
Another factor added uncertainty to these five alternative estimates: density. Until they get their hands on a sample, the team can only estimate a range of densities for Bennu’s regolith. The mass of the regolith they captured equals its volume times its density; less density, less mass, no matter the volume of the sample. A bag full of pumice is a lot lighter than the same bag full of marble. “It’s one of the areas where we could be surprised,” Moreau says, “and we could end up with a less massive sample if the density is much less than we assumed.”
In a meeting with NASA’s administrators, Moreau and his team used their five estimates to propose that the ship had collected several hundred grams of Bennu’s regolith, and perhaps much more. The team made the case that, despite the leakage, the Tagsam head still held well in excess of the minimum 60 grams, and they recommended stowing the sample immediately. The administrators approved, so the team sprinted to secure the head in its sample return capsule a week ahead of schedule. Unlike the touch-and-go maneuver, which was entirely automated, the stowing process involved manual visual checks and adjustments at every step of moving the arm into its secure position within the capsule. “It was a lot of work,” says Sierra Gonzales, a mission operations systems engineer who led the stow effort.
In practice maneuvers on the ground before the mission, the team had struggled to maneuver the arm into its secure lashings. Banging the capsule’s edge now, they worried, might spill regolith everywhere. But this time, the team aced the process at record speed—36 hours rather than four days. Moving the arm meant losing more asteroid dust, though. Imaging analysis showed that they again lost hundreds of particles, or tens of grams, during the maneuver. On October 27, a week after the Tagsam maneuver, the team did a pull test to verify that the head was locked into position, then fired pyro bolts to separate it from the arm and closed the sample return capsule. The remaining regolith was now locked in and ready to ship to Earth.
So how much space dust is now inside the return capsule? The team can’t offer a specific estimate, but one can imagine a few different scenarios, based on predictions about what might have happened during the Tagsam maneuver and the movements of the arm afterward. If the head came away from Bennu crammed full of material—say, a hulking 500 grams, which the team believes is within the realm of possibility—and leakage afterward during arm movements was in the low tens of grams, say 50 grams total, that would leave 450 grams of material in the capsule, more than seven times the mission’s required amount. The head could also have collected more like 200 grams of material, the low end of the team’s estimates, perhaps thanks to the surprising behavior of the asteroid’s viscous surface. Leakage in the tens of grams—again, let’s say 50 grams—would ensure a relatively good haul of 150 grams. But what if the ship lost more material than expected when the arm was moved? Doubling the leakage estimate to 100 grams of loss would cut the remaining amount of captured regolith in half. That’s still well within the mission’s parameters for success, but with a tighter margin for error if, say, the regolith’s density is surprisingly low.
Overall, Moreau says, it would take a significant disaster to end up with fewer than 60 grams stowed in the head: Some major leakage the team didn’t spot, a significant miscalculation about the amount of regolith they initially captured, an entirely missed dynamic or factor. “All our analysis says with more than 99 percent probability that we have at least 60 grams, and probably more than that,” Moreau says. Every team member WIRED spoke to sounded equally confident that their craft had stowed more—and maybe much more—than its 60-gram target.
If all goes well, in 2021, OSIRIS-REx will depart the vicinity of Bennu for Earth. In September 2023, the OSIRIS-REx will separate from the return capsule and steer itself to a lonely sleep somewhere in space. The capsule will fall through Earth’s atmosphere at 27,700 miles per hour, protected by a thick heat shield, and parachute to a landing zone at the Utah Test and Training Range. A retrieval team will track its location via optical techniques and radar. In the lab, researchers will finally be able to precisely measure the mass inside. Only then will the team know for sure how much space dust they’ve captured.