New evidence suggests that nuclear chain reactions at the Fukushima power plant didn't end when the generators were initially shut down, following the March 11 earthquake.
Technology Review explains why Tetsuo Matsui, a researcher at the University of Tokyo, thinks conditions inside two of the damaged generator units may have temporarily restarted chain reactions 12 days after the shutdown.
Nuclear reactors produce radioactive by-products that decay at different rates. One common by-product is iodine-131 which has a half life of about 8 days while another is cesium-137 with a half life of about 30 years.
When a reactor switches off, the iodine decays more quickly so the ratio between these two isotopes changes rapidly over a period of days. That's why measuring this ratio is a good way to work out when the nuclear reactions terminated.
The question on many people's minds is whether the hot nuclear fuel then melted allowing a critical mass of molten fuel to form, allowing chain reactions to restart.
Tetsuo Matsui at the University of Tokyo, says the limited data from Fukushima indicates that nuclear chain reactions must have reignited at Fuksuhima up to 12 days after the accident. Matsui says the evidence comes from measurements of the ratio of cesium-137 and iodine-131 at several points around the facility and in the seawater nearby. He has calculated what the starting ratio must have been by assuming the reactors had been operating for between 7 and 12 months.
He says the ratios from drains at reactors 1 and 3 at Fukushima are consistent with the nuclear reactions having terminated at the time of the earthquake. However, the data from the drain near reactor 2 and from the cooling pond at reactor 4, where spent fuel rods are stored, indicate that the reactions must have been burning much later.
READ MORE at Technology Review
EDIT: In the comments, David Voss makes a good point about this story, which I should have thought about. "Keep in mind that the Physics arxiv blog at Technology Review discusses preprints posted on the arxiv.org server. Arxiv.org is a valuable way for researchers to communicate, but these prepublication papers may or may not have been vetted by colleagues, may or may not have been peer-reviewed, may or may not eventually be published. There is a minimal level of filtering to remove overtly crazy stuff but that is about it. This may be relevant also to your post on "Is the rumor mill ruining particle physics?"
Via David Biello