Idaho Falls Read online

  Investigators were also told about the reactor's problem with boron. Designed to “poison” or dampen the reactivity of the uranium fuel, the boron had been flaking off the metal plate lining and sinking to the bottom of the reactor. That meant the distance the rods needed to be pulled to start a nuclear reaction was changing unpredictably. It was another intolerable condition that observers said would have led to the shutdown of a reactor that was better managed.

  SL-1 workers testified that supervisors, and even the AEC, which was supposedly monitoring the plant, had known about the sticking control rods and the loss of boron for months but had taken only piecemeal action to keep the reactor up and running. In fact, the official “fix” called for crewmen to lift and lower the control rods at the start of each shift until further notice. SL-1 supervisors said they thought that “exercising” the rods either with the drive mechanisms or by hand would keep the problem in check until a new reactor core was installed in the spring.

  Most disturbing of all, the committee became aware that even as the control rod problems increased, the level of supervision over the men working on the deteriorating equipment was reduced. In the first year and a half of the reactor's operation, three senior men were responsible for connecting or disconnecting the control rods. In the months leading up to the accident, that delicate work had been deemed “routine” and was transferred to regular military operating crews, which sometimes consisted of just two young men with perhaps a year's experience each. Committee members also discovered that the procedures for doing that work were detailed in no more than a couple of pages of broad instructions that contained less detail and warnings than those included with a gas barbecue.

  There was a feeling among some of the men assigned to SL-1 that the shutdown over the Christmas holidays had exacerbated the problems the plant had been experiencing. They suspected even more boron had been lost, raising the reactivity of the reactor further and radically lowering the number of inches a control rod needed to be raised to start a chain reaction in the core. They suspected conditions had deteriorated so drastically over the two-week shutdown that when a crewman lifted the central control rod just a few inches—either to latch it to the automatic drive mechanism, or maybe to “exercise” it—the reactor had gone critical.

  Richard Feil was an air force sergeant and chief reactor operator during the winter of 1961. He had clambered onto the top of the reactor the night of January 2 to disconnect the control rods so technicians could insert wires to test the condition of the core. More than forty years later, he counts himself lucky that he wasn't the one who ended up on the silo's ceiling.

  “The reactor was on the edge of stability, and it could have happened to anyone that night. It just depended on who lifted the rod,” he says. “The night before the accident, I had to pull the control rods up enough to disconnect them. It could have very easily happened to me—very easily.

  “We all knew that the boron was flaking off, but we didn't know how much,” he says. “Of course, the engineers I guess knew how much, but we weren't privy to that information. It took quite a bit of strength to lift that center rod and latch it onto the drive, and sometimes I had to jerk to get the initial movement. But I wouldn't call it ‘sticking.' It was just a matter of getting conditions right to lift the thing. The procedure was not at all unusual. I have no idea why the rod came up so far. I just know I lifted it up just enough to latch it onto the drive.

  “The accident was pretty sobering to us,” Feil recalls. “It was, ‘There but for the grace of God go I.'”

  Six days after the accident, the committee grilled army specialist Robert Meyer, a reactor operator at SL-1:

  Committee member: Is there anything you feel the [investigation] board should know, either related directly to the cause, or anything you might think might be of interest to the board, realizing that the board is here to find out as much as it can about the incident?

  Meyer: I'm not an engineer, but anytime you can grasp a rod and manually lift it is a poor business. That's what I felt when I first saw the arrangement they had here; this leaves it open for an incident of this sort.

  Committee member: But even so, you still think most people know this so well they would be careful in doing it?

  Meyer: That's right. But suppose this rod stuck a little bit; the rod is quite heavy, you know. It weighs around a hundred pounds with the shaft extension on it. Suppose it stuck a little bit and a man being slight, he probably hauled on it pretty hard. If this broke free, it's liable to come out quite a ways. This is one theory—it was sticking a little bit. We have had some trouble with the rods sticking.

  Meyer's testimony pointed to one possible series of events that would have culminated in an explosion just after 9 P.M. on that January night. Meyer's responses to the committee painted a picture: One of the three servicemen assigned to the central control rod had screwed a lifting tool onto its top. Bending his knees, the crewman had gauged the effort he'd need to raise the rod. It should not have required much effort, just enough to lift the nearly one-hundred-pound rod four inches so that another worker could attach a C-clamp to it near the base of the reactor top. The crewman pulled, but the rod wouldn't rise from the reactor core. Grasping the lifting tool tighter and sinking his knees a bit more for increased leverage, the crewman applied more pressure. The rod resisted the force, then suddenly broke free, hurtling up horrifyingly fast from the reactor core. Before the crewman could release pressure, the rod had traveled far enough out of the core—already unstable and dangerously reactive from the loss of boron—to cause an instantaneous explosion.

  Was this plausible? The committee members initially thought so. But as it turned out, there were a few details that couldn't be reconciled with such an accident scenario. First, the central control rod—the only rod that could have caused such devastation—was the only rod that didn't have a history of sticking. It had always slid in and out of the reactor core with ease, just as it should have. Second, the crew needed to lift the control rod only four inches to latch it to the drive mechanism. Yet the rod had to have been lifted at least sixteen inches to create the kind of prompt criticality that had occurred. Even if the rod stuck momentarily and the crewman maneuvering it applied some force, it seemed inconceivable that a fifteen-foot rod weighing nearly one hundred pounds would be withdrawn to that extreme. Third, supervisors told investigators that both Legg and Byrnes had reconnected the control rods at least four times each, and both had received extensive lecturing about never pulling a rod out too far or too quickly.

  There had to be something else that would explain the inexplicable. The transcripts of interviews with SL-1 workers run for hundreds of pages, and the thought processes of the committee members often have to be gleaned by inference and tone, by the kind of questions they were asking—or by the questions they weren't asking—and not necessarily by the answers they were getting. Still, it's clear that the panel was probing, ever so gently, into the qualifications and personalities of the crew on duty the night of January 3. Often, committee members prefaced their questions to SL-1 workers with phrases such as, “You know, you don't have to talk to us . . .” or, “If you'd rather not discuss this publicly . . .” Not surprisingly, those interviewed by the committee often took the out being offered to them, saying they couldn't really give an opinion on the performance of Byrnes or Legg. In some cases—presumably when a crew member was willing to talk about the two—the transcripts reveal the investigation committee would then quickly go off the record, breaking off a line of questioning about the competence and temperament of the men killed in the explosion.

  In the first week after the accident, the committee's interest focused on the trainee McKinley, whose body at that time was believed to be dangling precariously over the reactor. During Ed Vallario's appearance before the committee, the health physicist recounted the theory he had formed just hours after the explosion: “After understanding that McKinley was pinned to the ceiling, I felt that, knowin
g the operation, he had inadvertently withdrawn rod nine [the central control rod] in excess of fifteen inches and the result was an excursion.”

  The position of McKinley's body did seem to indicate that he was the crew member who lifted the rod. But that didn't make sense. He had been at SL-1 for only a month. He had never connected or disconnected control rods before, and other trainees were certainly not allowed to perform such a critical task. What was he doing up there on top of the reactor? The investigators began interrogating SL-1 supervisors and enlisted men about Legg, the supervising on-site operator for that evening's shift. Why would he let a trainee manipulate such a critical piece of equipment? Was he a good supervisor? Was he conscientious? How well trained was the sailor? Did he fool around—or did he let his crew do so—while on shift? The committee toyed with that possibility until January 12, when it learned from Lushbaugh's autopsy team that the bodies had originally been misidentified. As it turned out, McKinley was farthest from the reactor when it exploded, so it was unlikely he was manning the control rod. But that discovery, the committee realized, still didn't negate the theory that someone else on the crew had pulled the control rod too far, causing the explosion.

  By the end of January, the investigation committee had completed its interviews, and the members returned to their regular jobs on the East Coast. They would stay in touch with each other, meet in Washington, DC, when needed, and get regular briefings from the technical committee still working in Idaho Falls. They left quietly, and they weren't inclined to share their thoughts with the locals before they went. There were rumors that the committee had brought in special investigators to poke into the lives of Jack Byrnes and Dick Legg, but no one seemed to know what the easterners were looking for or even how serious or how far the sleuthing mission was being taken. Officially, committee members said, there wouldn't be a final report until the autopsy team and the technical advisory committee could reconstruct the crew and reactor's final moments.

  * * *

  In a flurry of long hours punctuated by frantic, minute-long forays into high radiation fields, the first phase of post-accident operations—the initial emergency response, retrieval of the bodies, and monitoring of the radiation levels in the environment—had taken just days. The second and third phase of activity—determining the stability of the damaged reactor and pinpointing the cause of the explosion through physical evidence recovered from the scene—would take considerably longer. Combustion Engineering would remain in charge of investigating the reactor's status. But it was clear that CE's role in managing the plant would be questioned during the investigation, so once the reactor was declared stable, the General Electric (GE) Company would take over gathering evidence from the site and then demolish the rest of the reactor.

  At the outset of February 1961, there was still long-life radiation zipping around inside the SL-1 reactor building, making the immediate area unsafe for the enormous group of personnel required for the final two phases of the operation. As a result, the checkpoint erected at Highway 20 and Fillmore Avenue on the night of the accident was expanded to become the official command center for the months of activity ahead.

  In addition to the decontamination trailers already in use at the checkpoint, two dressing trailers were hauled to the location. In the first, crews suited up in layers of protective clothing and donned respirators. Then they moved into the next trailer, the “Buffer Zone,” where their equipment was checked and individual radiation exposure badges were issued. From the second trailer, the crews emerged in the “Hot Zone” and were shuttled the one mile down the road to the SL-1 site. When their task at the reactor was completed, they returned to the checkpoint to begin decontamination procedures.

  Two other trailers that were brought to the checkpoint served as administration offices for CE and GE supervisors, and yet another became a makeshift electronics maintenance shop and communications center for keeping in contact with people working at the reactor.

  Inside the “Buffer Zone” trailer.

  With the command-and-control hub up and running, the search for water inside the reactor core began. Water plays a crucial, three-pronged role in a reactor's operations. First, it carries heat away from the fireball of fissioning uranium. With the SL-1 design, the water also boils, which creates the steam that turns the generators and produces electricity. Finally, water acts as a “moderator,” slowing down the dance of neutrons fired from the enriched uranium so that they slam into other atoms at the right speed to split off even more neutrons. The process continues until the core crackles into a self-sustaining nuclear reaction. If water was still trapped inside the damaged SL-1 reactor core, it was conceivable that the enriched uranium atoms left in the core could suddenly begin to fission. With the control rods completely destroyed by the explosion, there would be no way to control the reaction. But if the core was devoid of residual water, if it had all been spewed out in the violent initial blast, the reactor was likely stable.

  On the night following the accident, a first crude attempt was made to determine whether there was water inside the reactor. Resorting to primeval methods, the nuclear engineers and physicists instructed a soldier who was helping recover the second crewman to toss a rock into one of the open ports in the reactor top and listen for a splash. It was yet another unbelievable moment: cutting-edge nuclear technology reduced to a carnival pitch game. The soldier, hampered by a mask and time, couldn't manage to underhand a rock into the porthole. The scientists would have to devise another way to test for water.

  The team faced the perplexing problem of needing to survey the reactor without sending crews directly into the silo—the radiation levels were just too high. Using the full-scale mock-up of SL-1 that had helped crews plan the retrieval of the third crewman's body, the team practiced making remote entries into the reactor using TV cameras, drop lights, and a host of other specially adapted instruments. The equipment was attached to a twenty-five-foot traveling boom powered by a hydraulic crane. Throughout the months of February, March, and April, remote procedures were rehearsed and reviewed until every worker had perfected his role in the operation.

  By late April, the crews were ready. Each group received a final briefing before being transported down Fillmore Avenue to SL-1. The crane was positioned at the back of the reactor building and the boom, decorated with what was high-tech equipment at the time, entered the silo through the second-story freight doors, guided by an operator shielded in the crane's lead-lined cab. The TV cameras, also shielded by lead plates to prevent the film from fogging up in the high gamma-ray fields, revealed what many workers had caught only glimpses of in the Polaroid photographs taken right after the explosion: a violently twisted, viciously radioactive mess.

  When operators maneuvered the cameras through the open ports and into the guts of the reactor, the images beamed back seemed to show a core that was bone-dry. But the team needed to be absolutely sure. It performed two more remote penetrations of the reactor vessel. The first was with an ultrasonic vibration probe; when lowered to a depth only six inches above the top of the core, its frequency suggested that the vessel was dry. Those results were confirmed by a second probe, one covered with a water-soluble chemical. It returned from the bowels of SL-1 with its chemical component intact. There was no water left inside the reactor.

  With the core deemed stable, crews quickly shifted their focus to dismantling the reactor and discovering what secrets it held. That task fell to hundreds of GE employees, who a month earlier had faced unemployment when President Kennedy canceled their work on the air force's ill-conceived airplane reactor project at the site. The cancellation of the project meant that a massive hangar equipped to handle high-level radioactive waste was standing empty. Tearing down SL-1 would be tortuous work, but at least it was work. Because radiation levels were still so high, the time employees spent inside the reactor silo would be strictly limited, and the exposure spread out over a significant number of people.

  Performing a remote
entry.

  Beginning in June, a conga line of 1,240 people—GE employees, soldiers from Utah, and other site employees looking to pick up some overtime work—volunteered to take their quarterly maximum radiation dose. Decked out in canvas and plastic coveralls; surgeons' caps, hoods, and respirators; rubber boots, shoes covers, and cotton and plastic gloves to protect themselves from beta and alpha radiation, volunteers climbed the stairs to the second floor of the silo and did what they could in the two or three minutes they were allotted. Some loosened bolts just enough for the next workers to dash in and remove them. Others made short passes over the debris-littered floor with industrial vacuums before handing the nozzle to someone else. Health physicists stood outside at the bottom of the steps—eighty of them took rotating shifts to ensure their specialized skills wouldn't be lost because of overexposure—and banged on the railing when the workers' time was up. On average, it took workers four hours to suit up and strip down—all for those few fleeting moments of cleanup inside the reactor building. GE later estimated that throughout the operation, it used and disposed of approximately nine thousand pairs of shoe coverings.

  Bit by excruciating bit, the reactor floor was stripped of tools, rags, and small indistinguishable lumps that may or may not have been body parts. Pieces of the four control rod mechanisms ejected from the reactor were recovered, including the all-important central rod that had been lying across the top of the reactor. The debris was sealed in fifty-five-gallon drums and trucked miles across the desert to the giant hangar, called the TAN (Test Area North) Hot Shop. Recovered items that could potentially help solve the SL-1 puzzle were kept there; the rest were taken back to the SL-1 compound for disposal.