Taking toll at Tokai Mura

29 October 1999

Uncertainty remains over what exactly happened at Tokai Mura, but a picture is beginning to come into focus.

The criticality accident which took place at the Tokai Mura uranium processing plant on 30 September occurred in what was little more than a development rig, quite separate from the main uranium oxide production facility. The unit was built in 1984 and used for work on specialist fuels. In the words of a specialist who had visited the plant, it was operated “like a back-street workshop”. The accident appears to have happened as a result of extraordinary lapses in safety procedures and mismanagement by JCO, the company which ran the plant, driven by a desire to cut costs. It is likely that managers at JCO will face criminal prosecutions.

The process at Tokai Mura involves converting uranium hexafluoride into uranium oxide. In the process UF6 is first hydrolysed to remove the fluorine, and then the uranium product is then dissolved in nitric acid, forming uranium nitrate. This is then reacted with ammonia to precipitate uranium oxide.

The accident occurred during the process of converting uranium nitrate to uranium oxide. The plant is designed with three vessels for this part of the process, a dissolving tank, a storage tank and a precipitation vessel. The dissolving and storage tanks are designed to be ‘intrinsically safe’. They are geometrically designed to ensure a criticality reaction could not occur. The same is not true of the precipitation tank.

The approved process, with the uranium passing through each tank, is slow and JCO had revised its procedures to allow the uranium nitrate mixing to take place in 10 litre stainless steel buckets. The buckets’ contents were supposed to be poured into the dissolving tank and the maximum uranium content should have been limited to 2.4 kg. These changes to the process had been approved by a production manager and a managing director, but they had not been submitted to the regulators, because JCO knew they would not have been approved.

During last month’s accident, even these unsatisfactory controls were flouted.

JCO was making fuel for the fast breeder reactor at Joyo, hence the high level of uranium enrichment. The criticality occurred when two workers, Masato Shinohara and Hasashi Ouchi, poured seven buckets of uranium nitrate, containing 16 kg of uranium enriched to 18.8% U-235, into the precipitation vessel over a 12-hour period. On the evening of 29 September they finished the first stage of the process ahead of schedule and mixed four buckets of uranium nitrate containing 9.2 kg of uranium directly into the precipitation tank.

The next morning they prepared three more buckets containing 6.9 kg of uranium. At 10.35 am the contents of the seventh bucket entered the precipitation vessel, sparking the criticality. Shinohara and Ouchi reported seeing a blue flash and immediately felt ill. Both men, along with Yutaka Tokokawa, a third member of the team, were flown to the National Institute of Radiological Science. At 11.35 am the accident was reported to the local authorities and at 3.18 pm 160 people living within 350 metres of the plant were evacuated. At 10.30 pm the government advised 310 000 people living within a 10 km radius of the plant to stay indoors. Harvesting of agricultural produce was also banned.

JCO volunteers entered the plant at 2.35 am on 31 October to stop the reaction. Teams of two people each entered the contaminated area for a period of three minutes. By 4.19 am they had managed to drain water from the precipitation tank’s cooling jacket, eliminating neutron reflection, and had stopped the critical reaction. At 6.15 am the reaction was officially declared as over.

“The tank is almost cylindrical, with the bottom half a double structure surrounded by the cooling jacket,” said the Japan Atomic Industrial Forum (JAIF) in a statement after the accident. “It was thought that water originally intended to work as a coolant might be serving as a neutron reflector in the accident circumstances, facilitating criticality. A valve outside the building, in piping connected to the cooling jacket, was opened and the water was drained.” A criticality incident is usually a transient phenomenon. The solution heats up and the reaction stops, although it may go critical again when the liquid cools. But in this case radiation detectors and neutron detectors at the Japanese Atomic Energy Research Institute site 2 km away indicate that the reaction continued for 17 1/2 hours.

The neutron detector showed a large spike at the time of the accident, followed by slightly elevated readings for the remaining period. Occasional high gamma readings from fission products can be seen on the recorder chart when the wind was blowing from the JCO site.


Shinohara and Ouchi received very high radiation dose rates, initially thought to be 17 Gray and 10 Gray, respectively. However the 50% survival dose is 3-4 Gray and the fact that both men are still alive suggests they may not have received quite such large doses.

Both men remain extremely ill, with Hasashi Ouchi receiving transfusions of peripheral stem cells, which produce red blood corpuscles and white blood cells, from his brother, following a transfer to the Tokyo University Hospital.

Dr Robert Gale of the University of California at Los Angeles Medical Center, who treated some of the victims of the Chernobyl disaster, arrived in Japan on 8 October at the request of Tokyo University Hospital. He described Ouchi’s case as one of the worst he had ever seen. While supporting the work of the Japanese doctors, he said, “it is going to require a lot of work and a lot of good luck to get him through all of the problems he is going to face in the next few weeks”.

Japan’s Science and Technology Agency (STA) uprated the incident to level five on the International Nuclear Event Scale, after initially rating it at four. It is now on the same rating as the Three Mile Island accident in 1979, and only Chernobyl rates higher on the scale.

Initial assessments put the cause of the accident down to human error. The JAIF statement said: “There is a high likelihood of human error by a worker”. But the workers who carried out the procedure appear to have had almost no training, with one reportedly not even knowing what the term criticality means. The plant was regulated by the Science and Technology Agency and a picture is emerging of a belief within the STA that a criticality incident at the plant was an impossibility, which meant that planning for one was not necessary. Inspections were a bureaucratic formality, the workers had no protective clothing and were not even wearing film badges.

Police raided JCO headquarters on 6 October and rumours have been circulating widely since of the existence of a secret manual detailing procedures to cut corners and sidestep safety procedures. Legal proceedings based on violations of laws related to nuclear facilities as well as professional negligence resulting in injury are likely.

The corner cutting which appears to have been endemic at the plant reflects the increasing economic pressures JCO and other parts of the Japanese nuclear industry have felt over the last decade. According to the Wall Street Journal, in the early 1990s the Japanese public began to demand lower utility prices and in 1996 the government acted. The Ministry of International Trade and Industry forced power companies to cut rates by an average of 4.21%, with another cut of 4.67% following in 1998. To maintain profits power companies began looking overseas for cheaper nuclear fuel.

Between 1993 and 1998 JCO’s revenue fell by some 48%, with pre-tax profits falling from ¥400 million to ¥97 million.

It was also during the mid-1990s that JCO introduced changes aimed at simplifying fuel processing procedures, including the production of the “manual for standard procedures” which is now the centre of the police investigation.

Politics and Economics

There is likely to be considerable political and economic fallout from the accident, but despite its seriousness there has not been the type of public outcry in Japan that would be likely were a similar incident to occur in Europe or the US. This is despite the fact that Japan has 52 nuclear reactors in operation, and is highly dependent on nuclear power, as it lacks fossil fuels.

Sumitomo Metal Mining, JCO’s parent company, has already indicated it may withdraw from the nuclear business if the government permanently withdraws JCO’s operating licence. The Tokai Mura Society of Commerce and Industry (TSCI) has estimated the costs to local business at ¥300 million. The TSCI is planning to ask its member firms to estimate losses over a period of one month as part of a claim for compensation from Sumitomo. However, market analysts doubt the incident will have much impact on the credit quality of Japan’s nuclear generating companies.

A three member International Atomic Energy Agency team is investigating the incident and will report in mid-November.

The Japanese Labour Ministry has announced it will tighten its supervisory regime at nuclear installations, inspecting nuclear fuel processing plants and power stations every three months. In the past inspections have taken place once a year; in both the last two years the facility within the JCO plant where the criticality incident occurred was closed at the time of inspection. The ministry is also revising directives on nuclear processing facilities to make worker training obligatory.

The International Trade and Industry Ministry (MITI) has announced it will improve safety at nuclear power plants by introducing new monitoring and maintenance technology. Improvements include a system which uses sound waves to identify defective parts. MITI also plans to build a virtual plant where emergency situations can be simulated and drills conducted. The Ministry plans to contract Hitachi, Toshiba, Mitsubishi Heavy Industries and Mitsubishi Electric Corporation to develop the simulation technology, at a cost of ¥2 billion ($19 million) to the government.

The Federation of Electric Power Companies (FEPC) has decided to establish an organisation to exchange information within the nuclear generating companies. Modelled on the World Association of Nuclear Operators, it will have peer reviews, where members would visit each other’s facilities to promote best practice.

“We are determined to make every effort to restore people’s trust in nuclear safety,” said FEPC chairman Hiroji Ota. “Otherwise trust in nuclear power generation as a whole could be undermined.” Some 69 other people, including three ambulance workers, were irradiated during the incident.

Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.