Recounting the Windscale Story

The Blog, Histories of Things to Come, includes a post titled Nuclear Leaks 8, Sellafield, aka Windscale. The Nuclear Leaks part of the post title is the give away that the blog is not exactly cheer leading for nuclear power, but there is a good deal to learn here. Most of the story is told through links to a BBC documentary on the Windscale reactor fire. The BBC narrative fingers politicians, in particular British Prime Minister Harold MacMillan as playing a major causal role in the Windscale fire, by pressuring Windscale leadership to go beyond what was safe in reactor operations.

The original Windscale reactors were modeled after the ORNL graphite reactor, the world's first reactor designed to produce plutonium. The graphite reactor was low enough powered to be directly air cooled. That is air was physically blown through the reactor to remove heat. The Graphite reactor operated with natural uranium because enriched uranium was not available at the time it was built. With the uranium spread out through the graphite blocks in the reactor core, a chain reaction could be maintained. The British decided that the Graphite Reactor was safer than the water cooled Hanford Reactors. They wanted to produce nuclear bombs, so they decided to build Plutonium production reactors, using the reliable graphite pile approach.

The two Windscale reactors can be described as primitive. While the leaders of the American reactor building group at the University of Chicago, and in particular Eugene Wigner, had a very good theoretical understanding of what was happening inside a reactor, the British scientists who worked on the Windscale design seem to have been particularly remiss in the failure to appreciate the safety problems associated with this type of reactor. The British scientists had learned from the Americans that when bombarded by neutrons, relatively cool graphite could store a form of energy related to changes in its crystal structure. This energy, discovered by Eugene Wigner, and called Wigner energy, could be released by an increase in heat. The release of Wigner energy increased the heat contained in the nuclear core. In a large reactor the heat release could in turn trigger more Wigner energy releases in the core graphite, thus producing a sudden cascade of core heating.

The air cooled graphite piles operated at low temperatures - under 200 degrees Centigrade. Wigner energy could be released by increasing pile heat to over 250 degrees Centigrade.

Wigner also discovered Xenon-135 poisoning, which creates problems for reactor control. Xenon-135 related control problems can lead to uneven reactor heating and to overheating in some reactor areas. Xenon-135 atoms can put a break on chain reactions if present in large enough numbers in a reactor core. Increased power levels can in turn burn off Xenon and this can lead to a sudden surge in power output.

A further problem relates to the effect of heat on the Windscale fuel capsules. The Windscale reactor was fueled by metallic uranium, encapsulated in aluminum. Uranium swells when exposed to radiation or increased heat. That swelling in turn can burst the aluminum sheathing of the fuel capsule.

Metallic uranium is a pyrophoric substance. However uranium is more likely to burn if it has been powdered or shaved first. Lithium burns spontaneously if heated to about 356°F (180°C), and if Lithium is present in the core of an air cooled reactor it would be a huge fire hazard, especially in the presence of heating to release Wigner energy. If a core was heated to facilitate Wigner energy release, and a lithium containing capsule burst in it, the fire danger would be significant. The Windscale reactor cores housed capsules which contained both lithium and magnesium. Magnesium will not ignite in ordinary air, but it will catch on fire in heated air and of course a lithium fire will heat the air. The Penney Commission on the Windscale fire considered it possible that lithium initiated the fire, but appeared to consider uranium oxidation a more likely suspect. In their investigation of the Windscale fire, the Penney Commission found that capsules had burst by uranium expansion caused by radiation and heat. The expansion forced the aluminum top off the fuel capsule. Once the cap was removed the uranium top would be exposed to O2 in the air and would have begun to oxidize.

The Penney Commission did not focus on the oxidation of Windscale Unit 1 graphite during the October 1957 incident, although it gave a great deal of attention to the presumed release of Wigner energy from that graphite. Thus the Penney Commission was not in a position to determine the role of graphite fire in the Windscale accident.

Fred Pearce in The New Scientist asked,

What had gone wrong?

And answered,

the pile became too hot, cans of uranium split, the exposed uranium oxidized, releasing more heat, and eventually the graphite caught fire.

This is the classic graphite fire story. The story that has not been told in every recounting of the Windscale fire story. Where did the graphite fire come from? The New Scientist appears to have relentlessly pushed the graphite fire story, includes a 1982 account of a visit by Edward Teller to Harwell in 1948, Upon learning that the British were planning to build a graphite reactor, Teller is alleged to have warned British scientists of the dangers of graphite fires. Teller's warning was part of a standard account of the Windscale fire. No one has produced a contemporary account of Teller's Harwell visit, and the texts of the story are not sure whether or not Teller specifically mentioned a graphite fire, and indeed the operation of human memory as such, that Teller who was concerned about nuclear safety issues in general, did not warn of a reactor fire, but graphite popped into someones memory of the Teller warning after the Windscale event.

Remote inspections of the interior of the Windscale reactor do not demonstrate extensive evidence of graphite burning. The link is too a set of presentation graphics for the Windscale Piles Decommissioning Project. These graphics contain pictures of both damaged and undamaged fuel channels in Windscale Pile 1. Pictures of undamaged fuel capsule, fire damaged cartridges, all show graphite fuel channels which appear to be in good shape. Some oxidation may have taken place in the Pile 1 graphite, but the overall density loss of Pile 1 graphite is similar to the density loss of Pile 2 graphite. Thus conditions inside the reactor offers little evidence that a graphite fire took place, and indeed the fire damage to fuel capsules, suggest that they, and not the core graphite were the cause of the core fire.

The capsules containing lithium and magnesium were intended to produce tritium to be used in the British H bomb program. The decision to use the Windscale reactors for tritium production should have been vetoed for safety purposes. It wasn't. The British nuclear science community was shockingly unconcerned about nuclear safety issues. Not only was the community willing to allow the operation of reactors with unsafe materials in their cores, but they had no safety plans in the event of a major reactor accident.

On October 7, 1957 the Windscale staff began the annealing process expected to prevent the buildup of Wigner energy in Windscale Unit 1. The lithium-magnesium capsules were allowed to remain in the core during the annealing process the reactor despite the hazard they posed. The fans blowing cooling air into the core were turned down as was the reactor power. It was assumed that by decreasing cooling, the core temperature would rise to 250 degrees, enough to trigger a Wigner release. Once the Wigner release was assumed to have begun, the reactor was shut down. But the core heat was believed to drop off too quickly, evidence that the Wigner process had not been successful. At that point the Windscale staff decided to reheat the reactor. The second heating is believed to have some how triggered the fire.

It should not be said, however that the cause of the Windscale fire will never be known, because it would be possible to model the circumstances that gave rise to the fire, and test likely causes. At any rate the fire began to spread through the core as the aluminum sheathing of fuel capsules began to melt exposing more and more uranium to fire heated air. Aluminum is a class 4 flammable solid, and soon the molten aluminum began to add to the conflagration. It was only a matter of time before the fire was to spread to the uranium fuel.

The Windscale staff noted a rise in core temperature after the second attempt to accomplish a Wigner energy release, they tried to use passive cooling by opening core dampers, but the temperature continued to rise. Radioactive materials were observed flowing out the chimney stack.

At this point I will take a break for the day. There is a lot more to the story, and the story should be told. So look for more in the coming days.