Windscale Nuclear Incident


In 1957, the graphite moderator of one of the air-cooled plutonium production reactors at Windscale (now Sellafield), had a fire which resulted in the first significant release of radioactive material from a reactor. The reactor served a second purpose at the time - production of Po-210 (polonium) from bismuth. Po-210 was also released. These gas cooled reactors were operated by the British government at the time.

The map of England and Scotland below shows Sellafield in the center.

Courtesy BNFL

During the incident, radioactive releases included:

Radionuclide Estimated Range of Release (Curies)
I-131 16200 - 20000
Cs-137 600 - 1240
Sr-89 80-137
sr-90 6-9
Po-210 ~ 240

Reference: M. Eisenbud, Environmental Radioactivity (1987)

The fire started during the process of annealing the graphite structure. During normal operation, neutrons striking the graphite result in distortion of the crystal structure of the graphite. This distortion results in a buildup of stored energy in the graphite. The controlled heating annealing process was used to restore the graphite structure and release the stored energy. Unfortunately, in this case, excessive energy was released resulting in fuel damage. The metallic uranium fuel and the graphite then reacted with air and started burning.

The first indication of an abnormal condition was provided by air samplers about 1/2 mile away. Radioactivity levels were 10 times that normally found in air. Sampling closer to the reactor building confirmed radioactivity releases were occurring. Inspection of the core indicated the fuel elements in ~ 150 channels were overheated. After several hours of trying different methods to extinguish the fire, the reactor core was flooded with water. The plant was cooled down. Factors contributing to the event were:

  1. Inability to adequately monitor the core for damage

  2. Use of uranium metal, rather than uranium dioxide, as fuel. The metal has a lower melting point than the oxide.

Extensive sampling of milk, drinking water, and foods was conducted offsite following the event. The key radionuclides noted in the table above were analyzed.

Major conclusions were:

  1. The slow burning of the core resulted in the preferential release of radioactive iodine.

  2. The dose to any individual was greater from consumption of dairy products than from inhalation or direct exposure to the plume.

  3. Iodine contamination can be estimated from the gamma radiation levels in the area.

  4. If I-131 levels exceed 0.1 microcuries per liter in milk consumed, a child's thyroid dose could exceed 20 Rem.

Commercial reactors today use containments to reduce the likelihood of release. Uranium dioxide fuel is used in the majority of reactors. Emergency planning methods include better sampling and communication than in the 50's. The lower threshold for public concern would likely result in more restrictions on the use of dairy products.

Reference: M. Eisenbud, Environmental Radioactivity (1987)

Detailed Description

Most solids exposed to neutron radiation undergo change in physical properties. Graphite swells, its thermal and electrical conductivity decrease, and it tends to store thermal energy (also called Wigner Energy). By heating the graphite slowly, the Wigner energy can be released.

Graphite was used as the moderator in the core design of Windscale 1. Standard procedure after a normal reactor shutdown was to release the Wigner energy and restore the graphite to its original form. Eight such releases had been carried out successfully. The process is slow and time consuming.

On October 7, 1957 after a normal reactor shutdown, operators started the procedure to release the Wigner energy. The procedure involved starting the reactor at low power with the cooling blowers shut off. This heated the graphite and released the energy under controlled conditions. Following the first heat addition, operators noted the temperature was falling rather than rising. 

Subsequent investigation found that, in fact, in some parts of the reactor temperatures were decreasing, but a substantial number were increasing. This would indicate substantial areas were releasing energy, but some were not.

The next day, the operators added more power to release the Wigner energy. Because of a faulty power, the power was added too quickly. The temperature instruments were located in the positions of maximum temperature when the reactor was normally operating at full power. In this case, the maximum temperatures were elsewhere in the reactor. The releases of the Winger energy in the unmonitored areas got high enough to allow the graphite to catch fire. There were no unusual indication except for some variability in temperature. Steps were taken to cool and stabilize the reactor. 

On the fourth day, there were indications of radioactivity release through the offgas stack. Graphite temperatures also started to increase. Suspecting a ruptured fuel rod, the operators used a remote scanning device but the operating mechanism gear was jammed. Donning protective clothing, workers opened a plug on the front of the reactor and found the fuel was red hot. This was the fire indication of a fire that had been smoldering for about 2 days. Attempts were made to extinguish the fire which did not work. Finally on the 5th day, the reactor was flooded with water and the fire was extinguished.

The reactor was ruined. During the event, radioactive gases (primarily iodine and noble gases [krypton and xenon]) had been released. Meteorological conditions had varied throughout the event. Subsequent investigation showed that about 20,000 Curies of Iodine 131 had been released from the 405 foot stack.

Surveys in the surrounding countryside indicated the highest level had been at about 4 millirem per hour. Vegetative sampling indicated the stack filter had removed almost all of the radioactive particulates, but permitted the radioactive gases to be released. As a result, some gases as I-131 were transported to animal feed, which resulted in subsequent contamination of milk - the only effect on the public.

Radioactive analysis of milk over a larger area showed that the ban on milk distribution had to be extended to a total area of 200 square miles, beginning 2 or 3 miles north of the plant, extending over a strip 7 to 10 miles wide to a distance of 30 miles from the plant. The use of milk by the population in the restricted area was prohibited for 25 days. For the most highly contaminated areas, the prohibition was maintained for 44 days.

The Medical Research Council Committee concluded "that it is in the highest degree unlikely that any harm has been done to the health of anybody, whether a worker in the Windscale plant or a member of the general public." Except for the restrictions on milk usage, no other environmental action was required.

Descriptions of Selected Accidents that have occurred at Nuclear Reactor Facilities, H.W. Bertini and members of the staff of the Nuclear Safety Information Center, April 1980, pp 93-95

Additional Information

For additional information, contact the British agency responsible for nuclear safety and protection of the public from ionizing radiation , the Nuclear Safety Directorate, or the agency responsible for health and safety, Health and Safety Executive.

Copyright 1996-2006.  The Virtual Nuclear Tourist. All rights reserved. Revised: December 22, 2005.