NASA is about to make its most important safety decision in nearly a generation

As soon as this week, NASA officials will make perhaps the agency's most consequential safety decision in human spaceflight in 21 years.

NASA astronauts Butch Wilmore and Suni Williams are nearly 10 weeks into a test flight that was originally set to last a little more than one week. The two retired US Navy test pilots were the first people to fly into orbit on Boeing's Starliner spacecraft when it launched on June 5. Now, NASA officials aren't sure Starliner is safe enough to bring the astronauts home.

Three of the managers at the center of the pending decision, Ken Bowersox and Steve Stich from NASA and Boeing's LeRoy Cain, either had key roles in the ill-fated final flight of Space Shuttle Columbia in 2003 or felt the consequences of the accident.

At that time, officials misjudged the risk. Seven astronauts died, and the Space Shuttle Columbia was destroyed as it reentered the atmosphere over Texas. Bowersox, Stich, and Cain weren't the people making the call on the health of Columbia's heat shield in 2003, but they had front-row seats to the consequences.

Bowersox was an astronaut on the International Space Station when NASA lost Columbia. He and his crewmates were waiting to hitch a ride home on the next Space Shuttle mission, which was delayed two-and-a-half years in the wake of the Columbia accident. Instead, Bowersox's crew came back to Earth later that year on a Russian Soyuz capsule. After retiring from the astronaut corps, Bowersox worked at SpaceX and is now the head of NASA's spaceflight operations directorate.

Stich and Cain were NASA flight directors in 2003, and they remain well-respected in human spaceflight circles. Stich is now the manager of NASA's commercial crew program, and Cain is now a Boeing employee and chair of the company's Starliner mission director. For the ongoing Starliner mission, Bowersox, Stich, and Cain are in the decision-making chain.

All three joined NASA in the late 1980s, soon after the Challenger accident. They have seen NASA attempt to reshape its safety culture after both of NASA's fatal Space Shuttle tragedies. After Challenger, NASA's astronaut office had a more central role in safety decisions, and the agency made efforts to listen to dissent from engineers. Still, human flaws are inescapable, and NASA's culture was unable to alleviate them during Columbia's last flight in 2003.

NASA knew launching a Space Shuttle in cold weather reduced the safety margin on its solid rocket boosters, which led to the Challenger accident. And shuttle managers knew foam routinely fell off the external fuel tank. In a near-miss, one of these foam fragments hit a shuttle booster but didn't damage it, just two flights prior to Columbia's STS-107 mission.

"I have wondered if some in management roles today that were here when we lost Challenger and Columbia remember that in both of those tragedies, there were those that were not comfortable proceeding," Milt Heflin, a retired NASA flight director who spent 47 years at the agency, wrote in an email to Ars. "Today, those memories are still around."

"I suspect Stich and Cain are paying attention to the right stuff," Heflin wrote.

The question facing NASA's leadership today? Should the two astronauts return to Earth from the International Space Station in Boeing's Starliner spacecraft, with its history of thruster failures and helium leaks, or should they come home on a SpaceX Dragon capsule?

Under normal conditions, the first option is the choice everyone at NASA would like to make. It would be least disruptive to operations at the space station and would potentially maintain a clearer future for Boeing's Starliner program, which NASA would like to become operational for regular crew rotation flights to the station.

But some people at NASA aren't convinced this is the right call. Engineers still don't fully understand why five of the Starliner spacecraft's thrusters overheated and lost power as the capsule approached the space station for docking in June. Four of these five control jets are now back in action with near-normal performance, but managers would like to be sure the same thrusters—and maybe more—won't fail again as Starliner departs the station and heads for reentry.

The decision chain

A week and a half ago, this dissent in the halls of NASA caused managers to delay a high-level meeting to review the spacecraft's readiness to come home with Wilmore and Williams. NASA senior leaders would like to build unanimity among engineering teams before committing the astronauts to flying back to Earth on Boeing's crew capsule.

The latter option—to bring the Starliner astronauts home on a different spacecraft—is an alternative that was, at least on its face, unavailable to managers of the Space Shuttle program at the time of the Columbia accident. If Starliner comes back without its crew, this would almost certainly lead to debate within NASA about whether to require Boeing to complete yet another Starliner test flight before clearing the spacecraft for operational missions, as NASA cleared SpaceX to do in 2020.

Despite the pitfalls, many people at NASA believe this is the safer choice, although Boeing says it is confident in the Starliner spacecraft's ability to return the crew to Earth.

But ultimately, it's NASA's call. The lives of two government employees are in the balance, and taxpayers paid Boeing for most of the Starliner spacecraft's development costs. So far, NASA and Boeing have committed at least $6.7 billion to the program.

Officials believe they understand the cause of the helium leaks and have a plan to manage them on the flight back to Earth. But there are still uncertainties about the thrusters.

In a press briefing last week, Stich said NASA is making progress on a plan with SpaceX to return Wilmore and Williams on a Dragon spacecraft. Recent tests of a Starliner thruster at White Sands, New Mexico, produced some surprising results and left engineers still lacking an understanding of the fundamental cause of the overheating thrusters on Starliner in orbit. The majority view is that the overheating comes from rapid pulses of thrusters inside insulated doghouse-shaped propulsion pods on Starliner's service module.

Inspections of the thruster tested at White Sands showed bulging in a Teflon seal in an oxidizer valve known as a "poppet," which could restrict the flow of nitrogen tetroxide propellant. The thrusters consume the nitrogen tetroxide and mix it with hydrazine fuel for combustion. Despite the tests, however, engineers still don't understand precisely why the bulging is occurring and whether it will manifest on Starliner's flight back to Earth.

This discovery "upped the level of discomfort" among managers responsible for the Starliner test flight, Stich said.

While engineers continue assessing the thruster situation, NASA delayed the launch of the next SpaceX crew mission more than a month, to no earlier than September 24. This bought some extra time for NASA, although Stich said he would like the agency to make a decision by mid-August. That means a decision will likely come this week. Deciding now would allow time for SpaceX to reconfigure the internal cabin of the Dragon spacecraft for two astronauts rather than the normal complement of four crew members.

The scenario here would involve the Starliner astronauts staying at the space station until February, when the next SpaceX crew is slated to depart and come home. Wilmore and Williams would become fully integrated members of the station's long-term crew, taking the seats of two astronauts currently training to launch next month on SpaceX's Dragon.

If NASA goes this way, the Starliner spacecraft must undock from the space station without any astronauts aboard before the launch of the SpaceX crew mission next month. It will also take time for NASA and Boeing to update parameters in Starliner's flight software to enable an unpiloted undocking and reentry.

Questions about the end game for the Starliner test flight began circulating more widely at the end of July, when the thruster tests at White Sands proved inconclusive. In the weeks prior to the thruster tests, NASA and Boeing engineers worked on developing "flight rationale" to show that Starliner was acceptably safe to return home with Wilmore and Williams, despite the thruster problems and helium leaks.

"Good flight rationale describes the set of information that manifestly shows the part, subsystem, or operation will perform as required with requisite safety margins intact," wrote Wayne Hale, a retired NASA flight director, on his blog last week. "Alternatively, poor flight rationale includes unproven assumptions, incomplete testing, or analysis that contains flaws.

"Unfortunately, in the real world of spaceflight, there is seldom perfect flight rationale," Hale continued. "Sometimes there is good flight rationale. More often the rationale proposed for a flight contains ambiguities. Someone must exercise judgement to determine whether the flight rationale is adequate or not."

Last week, a meeting of key engineers at NASA known as the "Program Control Board" ended with no agreement on flight rationale for Starliner to return to Earth with its two-person crew.

"We heard from a lot of folks that had concerns," Bowersox said. "We heard enough voices that the decision was not clear at the Program Control Board."

Based on all that, Stich said the chances have increased that Starliner will come home without its crew, and Wilmore and Williams will instead return with SpaceX.

The Starliner decision will come down to senior leaders at NASA Headquarters. Normally, Bowersox, who chairs NASA's flight readiness review, would make the final call. He will hear briefings from NASA's engineering and safety teams, program managers, and representatives of the astronauts themselves. If officials present differing opinions to Bowersox, as they would have if the readiness review was held last week, the decision could go to NASA's most senior civil servant, Jim Free, or to NASA Administrator Bill Nelson, a former Florida senator who flew on the shuttle mission immediately preceding the Challenger accident.

"I chair that review, we'll poll everybody, see where we are," Bowersox said at a press conference Wednesday. "I'll state what my position is at that time. If we have dissenting voices out of that meeting, then it can go to the associate administrator, and if there's a desire to go even further, it can go to the administrator."

Last week, Nelson told Ars he has confidence in NASA's decision. "I especially have confidence since I have the final decision."

Familiar faces

The people in decision-making positions on the Starliner mission were all at NASA at the time of the Columbia accident.

Stich was a flight director on Columbia's final mission. He manned one of three shifts of flight control teams monitoring Columbia's activities in orbit on its 16-day research flight. NASA engineers, analyzing imagery from the Columbia's launch on January 16, 2003, discovered that a suitcase-sized piece of insulating foam fell from the shuttle's external fuel tank and struck a portion of the orbiter's left wing.

Engineers didn't know the damage a piece of relatively light foam could do to the shuttle's head shield when it struck the wing at more than 500 mph. Efforts to use spy satellites to take a picture of Columbia's heat shield were quashed by NASA management. At the time, NASA managers believed there was no way to rescue the crew, even if they found the heat shield catastrophically damaged.

NASA's flight directors, responsible for the minute-by-minute operation of Space Shuttle missions, got the decision on what to do about Columbia from higher-ups at NASA. Stich emailed Rick Husband, Columbia's commander, a week into the mission to inform him of the foam strike. "Experts have reviewed the high-speed photography, and there is no concern for RCC (Reinforced Carbon Carbon) or tile damage," Stich wrote. "We have seen this same phenomenon on several other flights, and there is absolutely no concern for entry."

Stich wrote that his reason for emailing Husband about the problem was to make the crew aware of it during an upcoming press conference. "This item is not even worth mentioning other than wanting to make sure that you are not surprised by it in a question from a reporter."

LeRoy Cain, now at Boeing, was the flight director on console in mission control at NASA's Johnson Space Center during Columbia's reentry on February 1, 2003. In real-time, over the course of about 10 minutes, Cain received updates from his team suggesting a cascading series of sensor failures on the shuttle, all clustered in Columbia's left wing. Then, mission control lost contact with the shuttle and its crew.

In an interview with Ars earlier this year, before the launch of the Starliner spacecraft, Cain said he would not hesitate to voice concerns about Starliner's safety. He is the top Boeing official in charge of day-to-day Starliner mission operations. "I would stand up and say we're not ready, and here's why," Cain said.

Cain presumably concurs with Boeing's official position that Starliner is capable of safely coming back to Earth with its two-person crew.

Bowersox was flying more than 200 miles above the planet, in command of the International Space Station. The grounding of NASA's Space Shuttle fleet after the loss of Columbia forced his crew to return home in a Soyuz capsule. And it was a wild ride. The capsule suffered a technical malfunction during reentry, causing its guidance system to revert to a steeper so-called "ballistic" descent, subjecting Bowersox's crew to higher-than-normal g-forces. They touched down nearly 300 miles short of their intended landing zone in Kazakhstan.

Known unknowns

There's one important distinction between NASA's decision about what to do with the Starliner spacecraft and the agency's mulling of the situation with Columbia more than 21 years ago.

Back then, the people in charge of the shuttle program weren't aware of the damage to the leading edge of Columbia's left wing. All of the data coming back from the shuttle showed it to be in excellent shape for reentry and landing. Today, managers know there are problems with Starliner.

"They frankly were fairly clueless on Columbia or certainly didn’t give it the same level of scrutiny, as they never thought foam shedding was an issue," a former NASA astronaut told Ars. "Also, on this one, Boeing is confident there isn’t a big issue with crew returning safely on Starliner, yet some within NASA certainly don’t feel that way."

Starliner's Reaction Control System (RCS) thrusters also had problems on an unpiloted test flight in 2022, and software problems caused Starliner's first test flight in 2019 to end prematurely. This "checkered history" could easily lead to some at NASA not having 100 percent confidence in the vehicle, the former astronaut said. "So there are some dissenters because there can be."

Scott Hubbard, a former director of NASA's Ames Research Center and a member of the Columbia Accident Investigation Board, said the "dearth of information" about the 10-inch hole in the shuttle's left wing led to a bad decision.

"In the current case with Starliner, much is known about the RCS (and) helium leaks, albeit not the root cause," Hubbard told Ars. "Based on public information, there are some tests of the valves which indicate that in an extreme case, there could be failures of the RCS system, which would be catastrophic."

Once the Starliner test flight is over, one way or another, Hubbard urged NASA and Boeing to revisit their decision to launch the spacecraft with a known helium leak. The craft's service module ended up leaking in five places. "Was the decision well-supported, or did someone have 'launch fever?'" Hubbard said.