The LANL review of criticality accidents is also interesting for the many, many ways to feed the fuckup fairy when it comes to this stuff.
There was a criticality due to contaminated vacuum pump oil (Yes, the mass in the oil was sufficient to take the container of pump oil critical).
One Soviet site is known to have had IIRC no less then three accidental criticality incidents in a YEAR, one of them in a rig intended to measure the chemistry to reduce the risk of such things...
Yes, duckies, when you change the geometry of the container holding the plutonium solution (say by tilting it) so that the stuff all runs to one side and the surface are a goes down, well, that would be known as an 'unfavourable configuration", it is generally a bad idea.
Yes, draining the vessel using a valve into 1L containers and moving them separately is slower then using a 25L drum, **BUT THERE WAS A REASON THE INSTRUCTIONS SAID TO DO IT THAT WAY**, some amazing stuff went on during the cold war on both sides of the curtain.
There is also a most amusing back and forth between the airfarce and the national lab concerning the cleanup after a fire in a missile silo that destroyed the warhead, with the airface trying to get the national lab to classify the concrete rubble based on its **AVERAGE** contamination, and the national lab pointing out that there would be bits **MUCH** hotter then average as something like 100g or so of special material was unaccounted for.
Some of the old public documents are fascinating.
Regards, Dan.