Water moderates, it doesn’t amplify. It slows the neutrons so that they’re better absorbed by the uranium, facilitating the chain reaction.
While the removal of the water would change the neutron absorption, the removal of the water would also allow the heat to build.
When the primary fission is stopped, by insertion of moderating rods, or whatever control mechanism is used, the production of heat does NOT stop. Many of the immediate daughter products of the reaction have a very short (hours) half life and continue to decay, making tremendous heat. You can shut down the reactor, but the heat continues to be produced, at very high levels for days afterwards (it’s a curve, but close enough).
That’s what happened at Fukushima - with a power loss, the moderating rods dropped fully into the reactor, absorbing neutrons and stopping uranium fission. As designed. A fail safe mechanism.
BUT - The power loss also caused the coolant circulation pumps to stop. And heat was still being produced. The water boiled off. The fuel rods melted and pooled at the bottom of the containment vessel and kept making heat, from decay, not necessarily from uranium fission, though that would increase as the fuel pooled, and they melted through the bottom of the reactor.
There are ways to make reactors safe, but in this design, you have to maintain coolant circulation for days after shutdown to manage the heat of decay.
Yes, I probably should have addressed that in my post too. A light water reactor uses regular water as a moderator. Other designs use different material, the CANDU uses deuterium (heavy water) which has a greater neutron slowing ability, allowing the use of natural (unenriched) uranium. Light water requires LEU in order to maintain fission, but military reactors used for propulsion use higher levels of enrichment to facilitate longer fuel cycles, reducing the frequency of refuelling events.
Another very effective moderator material is graphite, which also allows the use of natural uranium. The UK Magnox design is one such unit, which also does online refuelling like a CANDU, but probably the most famous is the Soviet RBMK design used at Chernobyl.
Control rods are used to, as the name implies, control the fission process, maintaining a certain level as well as reducing or increasing power levels by insertion or removal.
There are also shutdown rods which are only used to terminate the chain reaction, these are separate from the control rods. The CANDU 950 design for example had:
For reactor control:
- 27 stainless steel Adjuster rods
- 4 cadmium/stainless steel control absorbers
For reactor shutdown:
- 36 cadmium/stainless steel shut-off rods
- 8 horizontal nozzle tubes for liquid poison injection, which is gadolinium nitrate
At Fukushima, not only was there no way to power the pumps, due to the fuel tanks for the emergency generators being swept away, but from what I recall, the seawater cooling pumps were severely damaged as well, so even if there was power, there was no way to get cool water into the plant. On top of that, this caused issues with the used fuel storage pool, which they were having problems maintaining the water level in with no circulation or makeup water.
As you noted, the decay heat curve has a taper to it and its intensity is the highest right after fission is terminated. It takes around 8-10 years for decay heat to sufficiently decrease to a level where liquid cooled storage is no longer required and the bundles can be shuffled into dry casks. They of course still continue to make heat in the casks, but the level is quite low.
An interesting diagram of the heat transport system at Bruce: