Atomic accidents, p.34
Atomic Accidents, page 34
The tanks were part of the chemical separation system, meant to recover plutonium from machine-shop waste, leftovers in melting crucibles, or slag from casting. The tanks typically held aqueous solutions that were about 0.1 gram of plutonium per liter, which was way below anything that could be made critical, but the tanks, which had been in daily use for the past seven years, were obsolete, and they were scheduled to be replaced soon. They were still in fine condition, but they had been made in a perfect shape for accidental criticality. They had the surface-area-minimizing shape of a soup can, and the ends were rounded. By now it was realized that this was a dangerous shape, even though the procedures were designed to absolutely prohibit there ever being enough plutonium in one tank to go critical. The replacement tanks would be 10 feet high and six inches in diameter, which would discourage anything less than solid plutonium from becoming a runaway reactor.
It was 4:35 P.M. on December 30, 1958, a little before quitting time on the last shift before the New Year’s holiday. A load of 129 gallons of a murky fluid consisting of plutonium, nitric acid, water, and an organic solvent had been drained out of two other vessels and transferred to this particular tank.170 Allowed to sit for a while, the liquid had separated into 87 gallons of water in the bottom of the tank with 42 gallons of oily solvent sitting on top of the water. This was to be expected, which is why there was an aggressive stirring mechanism built into the tank. Unknown to anyone, plutonium solids, built up from years of processing, had dissolved off the insides of the tanks upstream and landed in this tank. The water in the bottom had only 2 ounces of plutonium dissolved in it, but the thin, disc-like layer of solvent on top contained a barely subcritical 6.8 pounds of plutonium, helped along in its quest to go critical by being homogeneously mixed with a hydrocarbon liquid, an excellent neutron moderator.
It was hard to predict this accident, except for the fact that the steel tank, built to use as little metal as possible, was an ideal shape for nuclear criticality using uranium dissolved in a liquid. When the electric motor was started to stir the two solutions together, it formed a whirlpool in the center. Instead of mixing immediately, the organic solution containing uranium was suddenly reduced in diameter and surrounded by water, making it a supercritical nuclear reactor.
Cecil Kelley had spent the last 11.5 years as a plutonium-process operator at Los Alamos, and he had almost seen it all. He stepped up on a footladder to look at the contents of the tank through a glass porthole on top, cupping his left hand to shut out ambient light. The ceiling fluorescents were illumining the surface of the liquid through another porthole on the other side. It was time to mix the water and the light solvent together. Leaning on the tank, he reached for the stir button with his right hand and pressed it. It took one second for the impeller to reach speed at 60 RPM.
A blast of heat washed over the front of Kelley’s body, going clean through him and coming out the back. It was like being in a microwave oven, as fast neutrons saturated his insides, exchanging momentum with his comparatively still hydrogen nuclei. He felt the strange tingling from gamma rays ionizing the sensitive nerve endings. A rushing noise was coming from the tank, over the whir of the stirring motor. Boiling? There was a slight tremor, moving the tank sideways very slightly, one centimeter, as it walked across the floor on its four legs. He fell backward onto the concrete floor. Dazed and confused, he got to his feet, turned off the stirrer with another push of the button, then turned it back on and ran out of the building.
Two other process workers in the same room saw a flash of blue light on the ceiling, as if a photoflash had gone off, and then they heard a dull thud. No criticality alarms went off, but they both knew that something bad had happened. They rushed to help Kelley and found him outside. He had lost control of his limbs. “I’m burning up!” he cried. “I’m burning up!” They hustled him to the emergency shower, turning off the stirrer as they passed it.
In a few minutes the medical emergency and radiation monitoring staff arrived. Kelley was in deep shock, phasing in and out of consciousness. He looked sunburned all over. By 4:53 they had him in the ambulance and headed for the lab hospital. The radiation monitors ran their Geiger counters over the tank. It was hot—tens of rads per hour. It was the remnant of a criticality in the tank, but how?
When Kelley started the stirring impeller at the bottom of the tank, it was supposed to mix the oily layer on top with the water on the bottom, and it would eventually do this, but first it started the water spinning in a circle, independent of the disc-shaped, plutonium-heavy solvent stratum. The water assumed the shape of a whirlpool, a cone-shaped depression in the middle of the tank. The solvent fell into the cone, losing its large surface area and becoming a shape favorable to fission with the neutron-reflective water surrounding it in a circle. Instantly the cone of solvent became prompt supercritical, releasing a blast of fast neutrons and gamma rays.171 The criticality only lasted for 0.2 seconds, but in that brief spike there were 1.5 3 1017 fissions. When the two fluids mixed together under the continued influence of the impeller, the plutonium-laden solvent was diluted by the water. The plutonium nuclei became too separated from one another for adequate neutron exchange, and the criticality died off as quickly as it had started.
Kelley’s condition was dire. He was semiconscious, retching, vomiting, and hyperventilating. His lips were blue, his skin was dusky red-violet, and his pulse and blood pressure were unobtainable. He was shaking, and his muscles were convulsing uncontrollably. His body was radioactive from neutron activation.
After an hour and forty minutes, he settled down and was perfectly coherent. He was moved to a private room. The staff drew blood and tried to get an estimate of his radiation dose from counting the activated sodium-24 in the sample. He had absorbed about 900 rad from fast neutrons and somewhere between 3,000 and 4,000 rad from gamma rays. A dose of 1,000 rad was thought fatal.172 His bone marrow had changed to inert, fatty tissue. He started having severe, uncontrollable pain in his abdomen, and he turned an ashen gray. At 35 hours after he had touched the stirrer switch, Cecil Kelley died.
The plutonium process was shut down for six weeks and the tanks were ripped out and replaced with the six-foot columns, as had been planned but put off.
Back in the 1960s, all the fuel reprocessing was not for weapons work, and all was not government-owned. There were also privately owned plants. The early startups were not large operations, but the ultimate goal was to take the spent fuel from power company reactors, extract the unused uranium, sell plutonium waste to the government, compact the fission products for efficient burial, and deprive the Canadians of a monopoly on the manufacture of medical isotopes, such as technetium-99M. Spent reactor fuel was seen as a cash cow, and not as a burden on the power industry. Fuel reprocessing was also considered a necessity for commercial breeder reactor operations, and breeders were expected to start coming online later in the decade.
Spent reactor fuel is mostly unused uranium, which is a natural part of the Earth’s crust and not particularly dangerous. If it could be processed down to the 1.4% that is highly radioactive, nuclear waste disposal would be much easier and reusable fuel would not be wasted.
The United Nuclear Corporation alone owned five plants. There was one in downtown New Haven, Connecticut, one in White Plains, New York, one in Hematite, Missouri, a research lab complete with a nuclear reactor in Pawling, New York, and a brand-new scrap uranium recovery facility in Wood River Junction, Rhode Island.
Before March 16, 1964, Wood River Junction was known for two things. It was the site of a railroad trestle washout leading to a passenger train disaster on April 19, 1873, and it was regarded the coldest spot in Rhode Island. On that day in March, the sparkling new United Nuclear Fuels Recovery Plant began operations. Its first contract was to recover highly enriched uranium from manufacturing scraps left on the floor at a government-owned fuel-element factory.173
The plant operated on 8-hour shifts, five days a week. Scrap material was received in 55-gallon drums as uranyl nitrate, diluted down to far-below-critical concentrations of uranium. The process to reduce the stuff down into uranium metal used the purex procedure. The incoming liquid was purified by mixing it with a solvent mixture of tributyl phosphate and kerosene, followed by adding nitric acid to strip out the uranium compounds. The uranium concentrate was then washed of kerosene residue using trichloroethylene, referred to as “TCE.”174
Robert Peabody, 37 years old, lived in nearby Charlestown with his wife, Anna, and their nine children, ranging from nearly 16 years to six months old, and he worked the second shift at the recovery plant. During the day, he worked as an auto mechanic, managing with two jobs to support his family. It was Friday, July 24, 1964, and Peabody had taken time off from his day job to go food shopping. It was getting close to 4:00 P.M. He dropped Ann and the dozen grocery bags at the house and took the five-minute drive to the plant.
The fuel recovery facility was a cluster of nondescript, windowless buildings, no more than three stories tall, painted a cheerful robin’s-egg blue. They were set far back from the road and surrounded by an imposing chain-link fence on a flat, 1,200-acre plot. Peabody clocked in, as usual, and changed into his coveralls.
It had been a nerve-racking week at the plant, mainly due to false criticality alarms. When processing nuclear materials, working with highly enriched uranium in aqueous solutions was about as dangerous as it could get, and an accidental criticality was something to be avoided at all costs. Wednesday he had been working on the second floor, washing down some equipment, when the criticality alarm sounded. Having it blare off nearby was like having a tooth drilled. He and the other four workers left the building in a hard sprint. It took a while to figure out that there was no danger, and that water had splashed into some electrical contacts. The radiation-detection equipment was set to be nervous and sensitive to the slightest provocation.
Later on the same shift, they discovered a substance technically designated “black goo” collecting at the end of the line where uranium was supposed to be coming out. The next day, the line had to be shut down, and everything had to be taken apart and cleaned, and the labeling conventions got a little scrambled as parts and wet rags and bottles were scattered around.
Most of chemical engineering practice had been fully automated by 1964, but this part of nuclear engineering, which was actually similar to other parts of nuclear engineering, seemed based in the nineteenth century. Processing nuclear fuel at this level was an astonishingly manual operation, requiring human beings to carry bottles of liquid material by hand and empty them into vats, tanks, or funnels at the tops of vertical pipes. The system was mostly gravity-driven, with liquid flowing naturally from the top floor to the ground floor in the plant. Somehow it seemed safer to have men carry around batches of uranium in small quantities, knowing exactly where and when it was to be transferred, than letting uncaring machinery do it.
The basic transfer units were specially built bottles, made of polyethylene, four feet tall and five inches in diameter, with plastic screw-on tops. Being tall and skinny, they discouraged a critical mass. No matter how concentrated was the uranium solution, it was impossible to get enough into one bottle to cause a chain reaction. However, it would be possible to stack them together in a corner and make a working nuclear reactor out of filled bottles, as stray neutrons flying back and forth from bottle to bottle would cross-connect them. That possibility was eliminated by having the anti-criticality bottles rolled around from place to place in “safe carts.” Each cart was built to hold the bottle vertically at the center of an open-framed, three-by-three-foot cube, made of angle irons and fitted with four casters at the bottom so it would roll. There was no way that all the bottles in the plant filled with highly enriched uranium solution could go critical as long as they were sitting in safe carts. The bottles were always separated by at least three feet of open space.
The problem with the polyethylene bottles was labeling them to identify the contents. Paper labels stuck on with Scotch tape would come off easily, because the bottles were frequently covered with slippery kerosene residue. The only way to get a label to stick was to hold it on with two rubber bands. On Thursday, the day shift was cleaning out the black goo in the system, and they found a plug of uranium nitrate crystals clogging a pipe. They cleaned it out with steam and drained the highly concentrated, bright yellow solution into polyethylene anti-criticality bottles. Paper labels were attached with rubber bands identifying them as containing a great deal of highly enriched uranium.
Five people at any one time ran the entire plant. On the night shift it was three young technicians, Peabody, George Spencer, and Robert Mastriani, the supervisor, a 30-year-old chemist named Clifford Smith, and the security guard. The plant superintendent, Richard Holthaus, was usually there during the day.
The back-breaking task of the evening was to clean the TCE, which had been used to wash the kerosene out of the uranium concentrate. It was expensive stuff, and it had to be recycled back into the process, but it always picked up a little uranium oxide when washing out the oil. The uranium was separated out of the TCE by adding some sodium carbonate and precipitating it to the bottom of the vessel. This process, like others in the plant, was carried out in small batches, and a shift-load of bottles loaded with dirty TCE was bunched up in safe carts. Peabody was expected to pick up each 35-pound bottle of solution, pour in some carbonate, and shake it for 20 minutes to ensure mixing. There had to be a better way.
Another way of agitating the TCE had been worked out in a previous shift. Weary of manipulating the heavy bottles, a technician had noticed that on the third floor was a perfectly good mixing bowl with a motorized stirrer, and why couldn’t we use that to slosh the TCE? It is not recorded, but I am sure he got the standard nuclear-work answer from the supervisor: “No! Give me a few minutes, and I will think of why you can’t do that.” Technicians could not be allowed to improve operating procedures on a whim. Eventually, the technician was able to wear down the supervisor, and word of an undocumented labor-saving procedure traveled through the plant with the speed of sound. The vessel in question, the carbonate make-up mixer, was about 18 inches in diameter and 26 inches tall, or the size and shape of a very efficient submarine reactor core. It was okay to use it, as long as the uranium content in what it was mixing was less than 800 parts per million, or very, very dilute.
It was nearly 6:00 P.M. Peabody rolled the safe cart with the first bottle in the cluster of what he assumed were bottles of TCE to be cleaned to the base of the stairs. The cart would not make it up the stairs, so he hefted the bottle to his shoulder. The label slipped out of the rubber bands and fluttered to the floor. The contents of this bottle looked about like the stuff in all the bottles. It was yellow, due to the extreme fluorescence of uranium salt, but this was not a bottle of contaminated TCE: it was uranium nitrate dissolved in water, from the black goo cleanout.
It was about as much work to get it up the stairs as it would be to shake the bottle, but Peabody arrived on the third floor, dragged the bottle over to the mixer, and unscrewed the top. The mixer was against the north wall of the room, held a couple of feet off the floor by metal legs, making the rim five feet high. The stirrer motor was hanging over the open top. Workers were protected from falling off the third floor and to the ground floor by a railing on either side of the narrow platform. Leaned against the railing on the right side, very near the mixer, was a folded two-section ladder, lying on its side.
The mixer already had 41 liters of sodium carbonate in it, and the motor was running. Peabody, who was only six inches taller than the lip of the vessel, stepped up on the sideways ladder and tilted the bottle into the mixer. Glug, glug, glug. As the last dregs emptied into the mixer, there was a bright blue flash and the sound of an enormous water balloon being slammed against the wall. As the geometry improved from long and thin to short and round, 6.2 pounds of nearly pure U-235 homogeneously mixed with water went prompt critical. Instantly, the contents of the mixer boiled violently, sending a vertical geyser hitting the ceiling, the walls, and thoroughly soaking Peabody with the products of 1 3 1017 fissions. He fell backwards off the ladder, jumped to his feet, and lunged for the stair well, screaming “Oh, my God!” The criticality alarm went off, and this time it meant it.
Peabody ran full tilt down the stairs, out the door, and was quickly making for the emergency shack, 450 feet away. His fellow workers were right behind him, fleeing the criticality alarm and watching Peabody tear his clothes off. He almost made it, but he fell to the ground naked, vomiting, and bleeding from the mouth and ears. Smith, the supervisor, ran to call Holthaus while Spencer and Mastriani grabbed a blanket from the shack and tried to wrap the injured man on the ground. He got up twice and tried to walk around, but he sank back to the ground with severe stomach cramps.
Soon the company officials and the police were backed up at the gate, and Peabody was loaded into the ambulance. His wife and eldest son, Charles “Chickie” Peabody, were found by a police officer. “There’s been an accident,” he began. “We’ll take you to the hospital.”
At 7:15 P.M., Richard Holthaus arrived at the plant, waving a radiation counter. Peabody had been the only person anywhere near the criticality, and he was the only one affected by the radiation burst. There was no radiation evidence on the ground floor that anything had happened, but nobody had turned off the criticality alarm, and the klaxon was still screaming. At 7:45 Smith joined him, and they cautiously climbed the staircase to the third floor, radiation probe held in front. There was no hint of a continuing criticality. Clearly, enough material had immediately boiled out to stop the chain reaction, but the walls, floor, and mixer showed fission-product contamination and were painted a brightly fluorescent yellow. Peabody’s bottle was still upended in the mixer. Holthaus went over to the mixer, removed the bottle, flipped the switch to turn off the stirrer, and quickly turned to go out the door. Smith took one last look and was right behind him. They had to quickly go downstairs and drain the contents of the mixer into anti-criticality bottles.
