The Engineer Who Spent Six Months Warning That Challenger’s O-Rings Would Fail in the Cold
The engineers’ charts were still on the table when senior vice president Jerry Mason turned to Bob Lund, Morton Thiokol’s vice president of engineering, and told him to take off his engineering hat and put on his management hat.
It was the night of January 27, 1986. Thiokol had just spent an hour telling NASA, over a conference line, that the next morning was too cold to launch the Space Shuttle. Now its own executives had muted the line, pulled into a private caucus, and the decision had come down to one man’s vote. Lund changed it. The four managers in the room signed off; the engineers were not asked. Thiokol reversed its recommendation and told NASA it was cleared to fly.
Roger Boisjoly, the seal expert who had spent the evening trying to stop exactly this, watched it happen and said nothing more. He had been making the case for six months.
What he knew
Boisjoly’s expertise was the seals. Thiokol built the solid rocket boosters for the shuttle, and the boosters came in segments, stacked and joined like sections of a chimney, each joint sealed by a pair of rubber O-rings meant to keep the burning gas inside. He had spent years studying those rings. He had concluded that in cold weather the rubber went stiff, lost its spring, and failed to close the gap fast enough at ignition. Hot gas could slip past.
In January 1985, a shuttle had flown after a cold Florida night, and when Boisjoly examined the recovered boosters he found something that frightened him: the primary O-ring on a field joint had eroded, with hot gas blowing past the seal. It was the clearest sign yet that the cold was the difference. Over the months that followed, further damage to the seals — a joint on a spring flight where the primary never sealed and the backup ring itself was eroded — convinced him the boosters were flying on borrowed time.
On July 31, 1985 — six months before Challenger — he put it in writing. In a memo to Bob Lund he laid out the erosion problem and the fear that it could appear in a field joint, where a leak would have nowhere to go. If that happened, he wrote, “the result would be a catastrophe of the highest order — loss of human life.” He wanted a dedicated team put on the seals. A small task force was eventually formed; it was given little support, and Boisjoly grew certain the problem was being managed rather than fixed.
The window closing
The forecast for the morning of January 28 called for an overnight low near 18 degrees Fahrenheit, with the pad still in the twenties by the planned launch time — far colder than any shuttle had ever flown, the coldest in the history of the program. Thiokol’s engineers asked for an emergency conference call with NASA. It began on the evening of the 27th, linking the company’s plant in Utah with NASA centers in Alabama and Florida.
Boisjoly and a colleague, Arnold Thompson, made the engineering case. No shuttle had ever flown with an O-ring colder than 53 degrees. The forecast had the pad in the low twenties before dawn, and even after weather delayed the liftoff two hours the air had only reached the mid-thirties — while the seal itself, cooled below the surrounding air, would be colder yet. Boisjoly spread a set of photographs across the table and walked through the erosion, the physics of stiff rubber and a joint that opened a hair at ignition. The Thiokol engineers recommended, unanimously, that NASA wait for warmer weather.
NASA pushed back hard. The shuttle was already badly behind schedule, and managers on the line were openly unhappy with the recommendation; one asked, in effect, whether Thiokol expected them to wait until April. Then Thiokol’s own executives asked for the private caucus — and the hat came off.
The refusal
Allan McDonald, Thiokol’s director of the booster project, was in Florida and would not sign the launch recommendation. If anything went wrong, he said, he did not want to be the man standing in front of a board of inquiry explaining why they had launched. The paper went up to NASA anyway, signed by another manager, Joe Kilminster. McDonald’s refusal went into the record and was overruled.
Boisjoly drove home that night sick with dread, certain the shuttle was going to blow up.
Seventy-three seconds
Challenger lifted off at 11:38 the next morning. Boisjoly watched on a monitor at the Utah plant, seated on the floor against the legs of another worried engineer, Bob Ebeling, who had a chair. As the shuttle cleared the tower and the seconds ticked past, the dread began to lift. Ebeling leaned down and whispered that they had dodged a bullet; a minute in, he said he had just finished a prayer of thanks.
Thirteen seconds later the screen filled with the white plume of the breakup. Challenger came apart 73 seconds into the flight, scattering across the sky over the Atlantic. All seven aboard were killed, among them Christa McAuliffe, the schoolteacher who was to have taught a lesson from orbit.
A cold-stiffened O-ring had failed in the right booster’s aft field joint, exactly where Boisjoly had said it would. Hot gas burned through and ignited the external fuel tank.
Afterward
When the Rogers Commission convened to investigate, Boisjoly and McDonald told it the truth: the engineers had recommended against the launch and been overruled the night before. Boisjoly called the closed caucus “an unethical decision-making forum resulting from intense customer intimidation.” Their testimony made the cover-up impossible.
It cost them. Both men were stripped of their old responsibilities at Thiokol, frozen out by colleagues who blamed them for the company’s disgrace. Congress intervened, threatening to bar Thiokol from federal contracts, and the company relented. McDonald was eventually put in charge of redesigning the very joint that had failed; he stayed until retirement and died in 2021 at 83.
Boisjoly never recovered the same way. He left the company, plagued by headaches and insomnia and depression, and built a second career lecturing engineering students on the ethics of speaking up — on what it costs to be the person in the room who says no and is right. He died in 2012 at 73.