The demon core was a ball of plutonium. It weighed about 6.2 kilograms. Scientists at Los Alamos in the United States used it for tests. In 1945, a heavy block fell from a scientist's hand onto the core.

In 1946, a tool slipped during a test. There was no explosion, but strong radiation came out. Two scientists died. After that, these tests were done by remote control.

The so-called demon core was a plutonium sphere of about 6.2 kilograms made at Los Alamos at the end of World War II. Scientists tested how close it was to criticality by surrounding it with materials that reflect neutrons back into it. In August 1945, a researcher was stacking heavy blocks around the core by hand when one slipped and fell onto it.

In May 1946, Louis Slotin held a reflector shell open with a screwdriver, and it slipped. In both cases there was no explosion, but a sudden burst of radiation killed the person working with the core. Afterwards, Los Alamos stopped doing such experiments by hand and switched to remote control.

The demon core, a nickel-plated plutonium sphere of about 6.2 kilograms produced at Los Alamos in the final days of World War II, earned its nickname only after two fatal criticality accidents. Researchers measured how near the core was to a sustained chain reaction by surrounding it with neutron reflectors. On 21 August 1945, Harry Daghlian was stacking 4.4-kilogram tungsten carbide blocks around it by hand when the final block slipped onto the core; he received about 510 rem and, according to the laboratory's official review, died 28 days later.

On 21 May 1946, Louis Slotin was demonstrating a procedure in which he held the upper half of a beryllium shell with his thumb while keeping a gap open with the tip of a screwdriver. The screwdriver slipped, the shell closed, and although he knocked it away at once, eight people were exposed, and Slotin died nine days later. Neither event caused an explosion; the harm came from an intense pulse of neutrons and gamma rays. Hands-on assembly was then replaced by remote operation. But the 1999 Tokaimura accident in Japan, partly blamed on a belief that criticality accidents could not happen, shows how easily such lessons fade.