The harvest in Buena Vista County began on an overcast day in September when the wind made the temperature feel like 38 degrees at eight o’clock in the morning. The sun broke through around noon, but then clouds moved in once again, and the temperature climbed to nearly 60 degrees by late afternoon. At that time of year the sun didn’t go down until almost eight o’clock in the evening, so the work could proceed apace, even with the occasional drizzling rain that blew across the fields in shifting layers of spectral mist. And yet by the end of the month, despite losing half an hour of daylight, the temperature soared to almost 83 degrees at midday under a cloudless sky. On those days, the sky was so blue that it reminded many people of the day in July when 1819 Uniform came plunging to earth and changed so many lives.
And indeed, as the harvest proceeded, the office of Chuck Eddy, the sheriff of Buena Vista County, had “taken on the look of an airplane salvage yard,” as Bill Zahren put it in the Sioux City Journal. Among the parts found since the crash were shredded aluminum scraps of the horizontal stabilizer, a four-foot section of the engine’s outer casing, parts of the cowling that had surrounded the missing fan disk, booster blades from the compressor stages of the engine, hydraulic lines—and enticingly, blades that had been installed on 00385, the missing fan disk. One young Iowa farm boy found an intact first-stage fan blade that was probably worth $1,000 in reward money. Instead of turning it in, he brought it to school for show-and-tell. On Tuesday, October 10, at around 3:30 in the afternoon, Janice Sorenson, fifty-eight, was running her combine through a field, harvesting corn near her farmhouse north of Alta, Iowa. Her family had been working that same 440-acre plot of ground since the late 1800s. The day was cool and pleasant, with clear skies and low humidity, and the workday had been slipping by uneventfully. Then she found that her combine wouldn’t move forward. “I felt resistance on the right-hand side, and it felt like something was stopping me. I backed up the combine and saw the fan blades sticking out of the ground. After I got out of the combine and got closer, I was really in shock.” She had no doubt about what it was. General Electric had stationed an investigator named Jerome Clark (no relation to John C.) in Buena Vista County, and he had aggressively promoted the cash rewards and circulated photographs of the missing pieces. In its fall from thirty-seven thousand feet the fan disk dished out and fell flat side to the earth, as John Clark had predicted it would. It was heavy, and it was going fast. When it hit the mud, it partially buried itself.
Janice Sorenson drove her combine the quarter mile to her home and called the sheriff’s office. Sheriff Eddy was in Des Moines. He left as soon as he received word that a piece of the disk had been found. He knew that no state trooper would try to stop him in his sheriff’s prowler, so he let it run wide open. When Eddy arrived at the Sorensons’ farm, “[The disk] was still there in the ground,” he said. “There was just about—I wouldn’t even say six inches of it sticking out of the ground. It was kind of spinning as it came down and just buried itself—threw up the dirt, and the dirt come right back down on top of it, and you couldn’t hardly see it.”
The disk was a dazzling silver as it lay in the cut-corn stubble. The blades were bent and torn and some were missing. Janice’s husband Dale helped to dig the disk out with a shovel. He drove his tractor over to the spot. He wrapped rags around the disk to avoid scratching it and put a chain through the bore. Then he lifted it with the bucket of his tractor and drove it to his turkey shed, where running hot water was available. With the disk hanging from the loader, Dale Sorenson sprayed off the mud with a hose. And even in its beat-up condition, it was, indeed, a beautiful object with its gleaming silver blades. In the bucolic farmland scene, it had the look of something that had traveled light years across the universe, a gift from an alien race in another galaxy.
As soon as Jerome Clark phoned General Electric, John Moehring called for a Lear jet to fly him to Iowa. The 406-pound piece of the fan, including disk and blades, was loaded onto that plane and rushed to Cincinnati, where by Wednesday night it was put under black security in Cell 10. Some of the old-timers at GE were able to take a look at it, and they knew what they were seeing. James Wildey, the senior metallurgist at NTSB, was chosen to lead the metallurgy group, including several experts from GE, one from United, along with a representative from the FAA. When he arrived at GE in the morning, Wildey walked into Cell 10 and saw the fan disk sitting on a skid on the floor. He walked up to it and peered at the fractured surface where about a third of the disk was missing. Already he could read any number of tales that the disk had to tell. For example, the number ten fan blade with all its fittings had been found at the Sioux City airport with its dovetail intact. That meant that as the disk cracked, the split had propagated right through the dovetail slot holding that blade. He had already seen the containment ring with seven evenly spaced witness marks showing where the fan blades hit. He guessed that those marks were made by the blades attached to the smaller piece of the fan disk, which was still missing. But most importantly, he read the cracked surface as only a trained metallurgist could.
“You could see that there was a brittle fracture region,” said Wildey, “which is typical of a fatigue crack. And that’s mainly what was visible on the fracture surface . . . the presence of a large fatigue crack that reached a critical size and broke through the disk. So that part was relatively straightforward. There was a preexisting crack in the disk, and then questions started to pop up: Well, why wasn’t this crack found? When did they have to inspect it?”
To get answers to those questions, Wildey and his team would ultimately have to destroy the very evidence that would prove their case. “We could see that there was a chipped-out region, a cavity at the bore surface.” The chipped-out region, the pit, was on the inside surface of the hole, or bore, in the center of the disk. The pit, or cavity, was measured by General Electric and found to be “0.86 inch aft from the forward bore face. . . . The dimensions were 0.015 inch radial depth, 0.055 inch axial and at least 0.030 inch in the circumferential direction of the bore,” according to GE’s own account. Although Wildey could see that the disk had cracked from the pit outward toward the fan blades, “at the time when we were first looking at it we didn’t know what exactly that meant.”
If the smaller third of the fan disk were found in a cornfield and if it contained more information, Wildey would be grateful. But it might never be found. What they had in hand might be all the information they ever received. Consequently, in testing, “you are very conservative,” said Wildey. “You try to get as much information as possible before you start doing destructive testing.” Destructive testing would ultimately involve grinding through the metal of that pit to find out what material was in there and why the pit existed where there should have been smooth and unblemished titanium alloy.
The first job was to get the fan disk into the lab and clean it. Even though Dale Sorenson had hosed it off, even though it looked clean to the naked eye, it had been in service for almost two decades, and then it had been sitting out in the mud and the rain for more than two months as helicopters and high-tech spy planes flew overhead and searchers tramped past it. Wildey wanted to look right into the heart of the metal. He did not want to be inadvertently looking at grease and mud or any stray organic material when he should have been looking at titanium or at whatever substance had made that metal fail. The first step was to use a brand new toothbrush designed for cleaning false teeth. Those toothbrushes can be dissolved by acetone, so the technician used soap and water and methanol to clean the crack, blowing the alcohol off with compressed air so that it didn’t evaporate and leave stains.
On Thursday, October 12, 1989, while Wildey’s team was working on the disk at the Evendale facility, Harold Halverson was disking a field of cut corn in Buena Vista County, running his tractor behind the combine that his son Allen was driving about half a mile east of where Janice Sorenson found the larger piece of 00385. It was about three o’clock in the afternoon, a sunny day with temperatures in the mid-70s. “The disk bounced,” said Allen, and his father stopped his tractor and stepped down onto the corn stubble. He slowly walked back to look. Before nightfall, the missing third of the fan disk was in the back of Jerome Clark’s station wagon. That same day General Electric delivered a check to Janice Sorenson for $116,000 for the larger piece of 00385 and the many parts and blades attached to it. You could have knocked her over with a feather.
The GE Lear jet flew out again, picked up the second piece of the disk, and flew it back, where it joined its mate in Cell 10. On October 16, it was sealed into a protective plastic bag and set aside untouched. Later that month, when asked under oath if the rewards program reached “a successful conclusion,” William Thompson said, “Yes, it did.”
But GE was not yet through with the search. By Friday, the company had organized more than a hundred people from metal detector clubs in Iowa, Nebraska, Minnesota, and Missouri. As Sheriff Eddy said, they “searched some fields for us.” Mick Erickson from Lincoln, Nebraska, found a bit of a fan blade. A few other bits of 1819 Uniform were found, but for the most part, said Eddy, “they found old harness buckles from back when they were plowing with horses, wrenches, nuts, bolts, and everything for machinery, screwdrivers, and all kinds of stuff out there.” On that farmland that had been worked for 150 years, they also found a jack knife, square nails, coins, a corn husking peg, and the bells from a genuine old-fashioned horse-drawn sleigh.
Employees from United Airlines took Susan White and Georgeann del Castillo to a dormitory at Briar Cliff College and sequestered them in a room. “I wanted to be out,” White said. “I wanted to be amongst everyone. [But] United supervisors wouldn’t allow us to mingle with the passengers. The only interaction we had with them was in the bathroom area.” White had intimately bonded with her passengers. She had Cynthia Muncey’s tears on her blouse, her sweat on her skin. White wondered where Cindy was now. The pathology report would list no injuries. It would say only, “Smoke inhalation (blood CO saturation 30%).” She had no idea what had become of Cinnamon, whose papers had been in her pocket. White had come all this way, and now she needed to complete her emotional journey with her passengers, whatever their fate, but United would not allow it. A woman from the Red Cross came to sit with the two flight attendants. The three women talked all night. White said that her ears and nose were clogged with Iowa loam, “and it kept coming out for weeks.”
The next day White’s father flew in. The NTSB interviewed her with representatives of the FAA, United Airlines, and her union attending. White was released, but she was afraid to get on a plane, glad that her father was there. Thursday night she and her father drove out of Sioux City and found a hotel. White could not turn off the lights in her room. “When I’d turn the light out, when I closed my eyes, I’d see just the fire and the bodies, and the crash, and I just—it was too terrifying for me, so I just lay there in the bed with the lights on all night.”
White said that she had been “just the happiest person in the world” before the crash. She loved people, and her outlook on the world was tirelessly bright and enthusiastic. Everyone loved her. But after three weeks of not sleeping, she was a wreck. She took a leave from work in order to concentrate on the therapy she received, but United Airlines kept calling her. “I couldn’t handle the pressure of them trying to get me to come back.” Reluctantly, she returned to work. Each time someone asked her about the crash, she would tell the story, but she would also suffer flashbacks and nightmares in her hotel room at night. Then she’d show up at work the next day exhausted, and the next crew would ask her to tell the story again. “So every time I’m at work I couldn’t escape it.” After a year of that, she suffered a breakdown in front of all the passengers and flight attendants on a trip. She returned to therapy and gradually put her emotions back together. After that she was just fine until the attack that destroyed the World Trade Center.
One of White’s good friends, Jason Dahl, was the captain of United Flight 93 on September 11, 2001, when it was hijacked and crashed near Shanksville, Pennsylvania. White was happily married by then, and she and her husband were supposed to leave for Greece that day. Instead they wound up at the Dahls’ house planning a memorial service. “And that’s when Two Thirty-Two really hit me. It hit me almost harder, or just as hard, as it did [in 1989]. And I didn’t realize that I hadn’t dealt with it. I still had issues. I had no idea.” After the attack on the World Trade Center, she wound up back in therapy. But she continued to fly, though not without difficulties.
“Anything unsettling on the airplane, severe turbulence, an approach [to] landing that doesn’t quite feel right, my palms sweat and I have horrible anxiety. No one would know it, but I feel it so intensely inside. Usually those nights when experiencing those feelings is when I’ll have a nightmare about crashing.”
In 1979, at the age of twenty-two, Nicholas Edward Cherolis graduated from college as a materials engineer. He was snapped up by General Electric, which had recently formed its first team devoted to nothing but analyzing the way metal parts fail in aircraft and engines. The group that Cherolis joined was a coterie of elite alchemists, sorcerers and their apprentices, who would delve into the heart of the crystalline structures inside of metals. They had an odd way of speaking, a language of their own, and when it came to metallurgy and failure analysis, they formed a closed and clubby society. Cherolis told me he’d “gone off and done a little fracture mechanics and fractography class with my buddy Doug Pridemore [of GE],” when the case of United Flight 232 came along. Ten years into his career, “it was perfect timing. I was totally prepared,” he said. The fan disk that Janice Sorenson and Harold Halverson dug out of the Iowa mud was a perfect match for Cherolis, who at thirty-two was reaching the peak of his skills. By then he was passionate about an obscure discipline called failure analysis.
When the fan disk arrived in Evendale, the team in the lab was busy examining the parts of the engine that had already come in. The smear tests, done by Joe Epperson at the NTSB labs in Washington, had already given hard evidence that fragments from the number one fan had cut the number one and number three hydraulic lines. When the electron beam went down into the inner electron shells of those atoms and found titanium, the team knew that metal could have come only from fragments of the fan disk or its blades. Now the team had the disk in hand and had to dig down to the final level and show why the disk broke apart in the first place.
“The most likely thing is that it has split from a fatigue crack,” Cherolis said. “The old guy in the group gets out his pictures and says, ‘See, here’s the history,’ and he had a cross section of an engine with little circles showing past failures of the same sort.” As a metallurgist in the field of failure analysis at a company that made turbine engines, Cherolis knew that he would eventually be called upon to analyze titanium that had been tainted. It was a matter of letting the wheels spin and being patient. Cherolis also knew that disk 00385 could have been damaged in handling or that the stresses induced in the disk during manufacture might not have been properly accounted for. “But by that time this engine series was well wrung out,” he said, “and that possibility was pretty slim. There were still problems in the turbine [of the CF6-6], but the fan hadn’t had any stress problems.”
Cherolis first looked at the larger piece of the disk on Thursday morning, October 11. He said, “You could see it [the crack] had come from the bore area. You can read a fracture backwards to where it starts. And it doesn’t look like a normal fracture that’s purely fatigue. There’s a depression of missing material at the origin that a layman would call a pit.” The crack had begun in that pit, the chipped-out cavity that Wizniak had called the origin. Cherolis expected to see smooth, continuous machined metal on the inside surface of the bore. He saw instead that tiny hole where a piece of material had fallen out. To Cherolis’s eye, it was evidence as damning as a bloody thumbprint at a murder scene.
Once the disk was cleaned, Cherolis and his team began taking photographs of that surface and the crack that grew from it, all the way to the rim of the disk. First he mounted a Polaroid camera on a tripod to take overall photos. Then he used a device called a macroscope to take photographs at five to ten times magnification. The team established X and Y coordinates on the fractured metal so that each photograph could be matched with others to orient all the features in space. This also allowed them to determine the direction in which the fracture was propagating and how fast it had been growing.
Wildey, leading the team, could see how the cracked surface changed as it moved out from the bore. The crack clearly began as a fatigue crack, but then it abruptly changed to what Wildey called “single load,” something that was ripped apart by brute force, or in the language of metallurgy, “with a single application of load.” The fatigue area nearer the bore looked completely different as the result of repeated cracking by small increments each time a load was put on the part. The word load means pull or force. The pull comes from centrifugal force, the force that results from spinning. Usually such a fatigue crack would let go on takeoff because that’s when the load is greatest. (And this is what the engineers hope for, since the pilot can usually abort the takeoff.)
Above that fatigue crack, along the edge of the single-load cracking, Wildey could see a shear lip, a flap of metal that droops off of one side at about a 45-degree angle when metal is torn apart. He said that when the piece pulled apart on the day of the crash, it was “like Thor smiting it with his hammer. This is a big deal. This is the Big Bang that everybody hears when this thing breaks. This is the clap of thunder” heard by farmers on the ground. “You’re at a stress level that’s exceedingly high and most often you’re just ripping the part apart and it’s all done. But because we don’t have that shear lip down there in that fracture region, it says that there’s some kind of brittle fracture mechanism that’s occurring down there. An experienced fractographer would be drawn to that area without a shear lip instantaneously.” And he would say, “There is a lack of deformation going on here that indicates for sure that there is a brittle fracture mechanism.” Wildey knew. Cherolis knew. Now the team had to prove it.
As the day wore on, Cherolis viewed the small pit inside the bore at higher and higher magnifications, taking photos at each step. “And you compare the surface finish away from it to the surface finish in it,” he said. On the healthy metal surrounding the pit, he could see the dimpled appearance that resulted from shot peening. That process gives a texture to the titanium that is almost like skin. It compresses the metal near the surface and thereby introduces stresses that help prevent cracks. But down in the pit, at higher magnifications, he could see two things of great interest. One was evidence of shot peening inside the pit. That meant that before the disk was completely finished, a bit of material had already fallen out to create the pit. Moreover, inside the pit he could see that the shot peening had fractured the metal further, which told him that the material was, indeed, brittle in there. “It looked like little pieces were missing out of that,” Cherolis said. “Well, it doesn’t do that on nice ductile material.” Healthy titanium alloy should not fracture under the force of shot peening. So the material in the pit had not behaved the way titanium should. Titanium should stretch under the load of spinning. The material in the pit, almost ceramic in its consistency, would not stretch. “The GE guys really don’t like to admit that there could be micro-cracks within the [pit],” said Cherolis.
United Airlines, for its part, denied that the cavity existed in a detectable form before the first flight. In its official report United said, “At all times prior to the inflight event, the cavity was filled with metal and was not visually detectable.” In fact, the major parties to the event wrote their own reports detailing what they wanted the NTSB to conclude. General Electric went so far as to print its report in the exact same format as the official NTSB report, including the same typefaces, so that the GE report could easily be mistaken for a genuine NTSB report.
In terms of whether or not the defect was detectable, Cherolis said that during the machining process, tools would cut metal from every surface to achieve the final shape of the disk, and when a tool hit the brittle material in the pit, “it was like hitting a rock in peanut butter.” His conclusion: “In my mind it was probably already micro-cracked” before it was ever installed on an engine. Tellingly, electron micrographs would clearly show cracks in the micro-structure, “and not all in the direction that the stress [of spinning] would make them.” There are “extra little cracks in random directions.” In a sense, it makes no difference who’s wrong and who’s right. If the defect hadn’t begun to crack during manufacture, it would have cracked the first time the engine was spun up to speed.
In its final report, the NTSB put it this way:
The Safety Board believes that at the time of manufacture of the disk, the cavity at the fatigue origin point was originally filled, or nearly filled . . . making the defect more difficult to detect. . . . The cavity was most likely created during the final machining and/or shot peening process and . . . the shot peening probably created the microcracking parallel to and just below the cavity surface. Moreover, the shot peening quite likely created the mechanical deformation on portions of the cavity bottom.
After that, every time an engine bearing fan disk 00385 was started, the crack grew a bit more, elongating outward from the bore toward the rim and the dovetail slot where the number 10 blade was attached. When that fatigue crack had grown to about one inch long and half an inch deep, the disk was ready to let go on the next flight. That flight took place July 19, 1989. You can see it in the photographs: the fatigue crack is flat where the metal broke as a crack through a broken plate of glass would be. But when the disk let go, the crack propagated and tore through healthy titanium as if through living flesh, leaving ragged edges. The fracture traveled through the metal at the speed of sound.
“It’s trying to fly apart all the time,” Cherolis said, “and the whole art of making a jet engine is keeping it all together and figuring out how long your parts will last. And you actually retire them with no cracks in them and throw them away, because statistically one could start cracking some time in the next several hundred cycles. And without a defect, that all works wonderfully. But you put a defect in there, then the crack starts right away, and all your calculations are trying to avoid starting a crack. And here it’s already there.”
Wildey and Cherolis now had to answer the next question. They knew how the disk broke. It had something brittle in that little pit, something that cracked and fell out, destroying the integrity of the titanium matrix. So what was that something? By the time Cherolis finished his work, it was late at night. Someone at GE telephoned Floyd Brate at about midnight and woke him up, saying he’d better get to the lab now. He was an expert in making detailed replicas of cracks. Other technicians were called in too. It was time to find out what material lay inside the pit.