The operator who had made the first criticality suffered from acute, severe radiation sickness. He had absorbed 700 rems of mixed radiation. His vascular system collapsed, and eventually one hand and both legs had to be amputated, but he was still alive 34 years after the accident. A little over a month after the accident, the shift supervisor died.
Mayak is still in business, and safety conditions improved over the decades from “medieval” to levels in keeping with 21st-century handling of radioactive and potentially critical materials. There has not been a criticality incident since the fatal accident in 1968, and the last death at the plant was in 1990, caused by a chemical explosion in a reagent tank.
The last fatal accident due to an unplanned criticality occurred in Japan in 1999, in a publicly owned nuclear-fuel-processing plant. This accident was unusual in that the criticality was not over in a flash, but would continue to react for an impressive 20 hours, and the two men who died broke the records for length of survival after receiving lethal radiation dosage. It was similar to the previous accident in Rhode Island, in that a break with the standard procedure to make the work easier led to the criticality, and even in 1999 the fuel processing incorporated a surprising amount of manual labor. It was also the first and only criticality accident in which members of the public not involved with uranium processing were exposed to measurable radiation.
The Japan Nuclear Fuel Conversion Co. Ltd. was established in 1979 as a subsidiary of the Sumitomo Metal Mining Co. Ltd. The Fuel Fabrication plant was built in Tokaimura, Ibarakin Prefecture, Japan, on a 37-acre, inner-city plot of ground. Unlike the United States or Russia, where a nuclear plant of any purpose was built in a lonely, isolated place, in Japan it was put in a highly congested, tightly packed city of over 35,000 people. In two large buildings, incoming source material, uranium hexafluoride gas, was converted to either uranium oxide powder or uranyl nitrate dissolved in water. The plant handled uranium used in light-water commercial power reactors. It was a large-scale plant, handling 540 tons of uranium per year at the peak in 1993, but it was only licensed to process low-enriched fuel, about five percent U-235.[181] Competition with foreign companies doing the same thing was stiff and production efficiency always needed tightening, but in 1993 the company sold ¥3,276,000,000, or $32,760,000, worth of product.
In 1983, a small facility, the Fuel Conversion Test Building, was erected to be used for special products. The plant’s license was modified to allow the processing of uranium enriched to up to 20 percent U-235 so that startup fuel for the Jōyō fast breeder reactor could be produced. Jōyō needed fuel enriched to 18.8 percent U-235. Care was supposedly taken in the building’s design to ensure that no enriched uranium would ever be in a critical-sized or — shaped container, so no criticality alarms were called for in the license. An accidental criticality of any kind in this facility, run by highly disciplined Japanese laborers, was not a credible scenario.[182] Gamma-ray detectors were bolted to the walls in all the buildings, in case some mildly radioactive fuel was somehow misplaced.
A step in the licensed procedure for making highly enriched uranyl nitrate was to mix uranium oxide powder and nitric acid together in a dissolver tank. As the nitrate product dripped through the dissolver, it was conveyed by a stainless steel pipe to a long, thin stainless steel holding vessel, specifically designed not to allow a critical mass of liquefied uranium solution to exist in it. The uranyl nitrate solution was then drained out the bottom of the vessel into small polyethylene bottles, each holding a non-critical four liters of solution. A little petcock on the bottom of the vessel controlled the flow into a bottle held under it. Just follow the procedure, being careful not to stack the bottles close together, of course, and nothing can happen.
In 1998 the company’s name was shortened to JCO, requiring less ink to print. By then the fuel-conversion business had fallen to 53 percent of the peak back in 1993, but in September 1999 JCO won a contract to convert 16.8 kilograms of uranium into uranyl nitrate for Jōyō. On September 29, three operators, Masato Shinohara, Yutaka Yokokawa, and Hisashi Ouchi, were assigned the task of dissolving the uranium oxide in nitric acid in seven batches of 2.4 kilograms of uranium each. With each run of uranium being only 2.4 kilograms, there was no chance of criticality.
There was an immediate problem. The drain petcock on the bottom of the long, thin holding vessel was only four inches off the floor. There was no way to fit a bottle under it. The resourceful workers decided to mix the uranium oxide with acid in a 10-liter stainless steel bucket instead. They could then tip the bucket, pour the solution into a five-liter glass Erlenmeyer flask, and then dump it directly into Precipitation Tank B, which had an electrically driven stirrer. This would save time by not having the solution sit around in little four-liter bottles, and the stirrer in Tank B would do the job a lot faster than just letting it drip through the dissolver. This plan indicated a weak understanding of the factors that lead to criticality. True, 45 liters of 18.8-percent enriched uranium solution is not critical, but only if it is in a geometry that does not encourage criticality, such as the long, thin tank. The 100-liter Precipitation Tank B was round and short, meant to incorporate as little expensive stainless steel as possible in its design, and it was therefore an ideal reactor vessel.
Ouchi stood on the metal platform surrounding the top of the tank, holding a glass funnel with his body draped over it. Shinohara climbed the metal steps to the platform, carefully cradling the flask full of solution, and poured it slowly into the funnel. Yokokawa sat at a desk nearby and completed the paperwork. By quitting time, they had successfully processed four batches, now sitting in Precipitation Tank B.
Next morning, it was more of the same. By 10:35 a.m. they had done two more batches, and they were almost through pouring the last of batch number seven into the tank. There were 0.183 liters left in the flask. Drip. Drip. There was a blue flash out the open port, lighting up the ceiling. Shinohara and Ouchi staggered down the steps, starting to feel strange. Then came extreme abdominal pain, waves of nausea, and difficulty taking a breath. Yokogawa looked up from his paperwork and turned in his seat, quizzical. The three workers had no idea what had happened, but the gamma-ray alarms were sounding. Ouchi had lost control of his muscles and was sinking into incoherence. His two fellow workers helped him out of the building. Someone had released gamma radiation somewhere in the plant, and they had to get out of the building. The unshielded reactor they had assembled in Precipitation Tank B was still running at power, boiling the uranium solution and broadcasting a deadly mix of gamma rays and neutrons in all directions.
Workers in all three buildings were streaming out and going to the emergency mustering point as the gamma-ray alarms rang everywhere. A worker from the building next door noticed that three guys from the Fuel Conversion Test Building looked injured and confused. He summoned an ambulance, and they were quickly removed to the nearest hospital.
181
The company wished to process Mixed OXide fuel (MOX), which is a combination of uranium and plutonium, derived from reprocessing spent power-reactor fuel. The United States opposed this plan under a new nuclear nonproliferation policy, fearing that Japan would either secretly build up a nuclear weapons stockpile or sell plutonium under the table to some other Asian tiger, neither of which seemed likely. Heated negotiations went on for three years, beginning in 1977. The U.S. finally gave in, agreeing to a proliferation-resistant process for mixing plutonium and uranium devised by the Japanese Power Reactor and Nuclear Fuel Development Corporation (PNC). Several power reactors, including Fukushima Daiichi 6, have been operated using MOX fuel, saving uranium and burning off otherwise unusable plutonium.
182
This “credible accident” criterion would bedevil the Japanese nuclear industry on March 11, 2011, when the Fukushima Daiichi power plant was knocked out of service by an earthquake and tsunami wave. This was not a credible scenario, so no preparations were made to prevent damage during such an event.