All bad nuclear movies have that scene where someone shouts “God help us, the reactor is critical!” Every nuclear trained person’s favorite bit of pedantry is to correct people who use the term critical to describe an out of control reactor.
Actually, a critical reactor is one where the neutron chain reaction is self sustaining. This means that for every generation of fission, there are enough neutrons flying around the inside of the reactor to make exactly that many fissions happen again in the next generation.
For a very simplified math problem (these numbers are all off by several orders of magnitude), let’s say 100 neutrons are in the reactor. They go on to cause 50 fissions to happen. Those 50 fissions release 100 more neutrons, which go on to cause 50 more fissions. Repeat forever. The amount of fissions per second happening in the reactor is constant at 50, so reactor power is constant. A critical reactor is a reactor with a constant power.
A reactor can be critical, sub-critical, or super-critical. Sub-critical means that you lose some neutrons each generation, so fewer fissions happen each generation. Reactor power is going down. A super-critical reactor is the opposite. Each generation produces more neutrons than the last one, which makes more fissions happen, which makes more neutrons, .etc. In a super-critical reactor, power is rising.
There is one more kind of criticality. Prompt critical. This is what happened at SL-1 on this date in 1961. SL-1 was a military test reactor. It had been shutdown over the holidays, and the crew was preparing to restart the reactor. To do this, they needed to hook the control rods to the mechanism used to move them. This required pulling the control rod up by hand a few inches.
Unfortunately, the technician attaching the control rod was a typical guy and wasn’t sure what 4 inches was. He somehow managed to pull it out about 2 feet instead.
Control rods absorb neutrons. This prevents the neutrons from reaching the U-235 fuel. The rush of neutrons that were suddenly free to hit the fuel when he pulled the control rod too far was sufficient to drive the reactor very, very super-critical. This is where it gets a little more complicated.
There are a lot of different ways to talk about the neutrons in a reactor. The key to this discussion is prompt neutrons vs delayed neutrons. Prompt neutrons are released in the instant the fission occurs (within .00000000000001 seconds.) The vast majority (99.9%) of all neutrons in the reactor are prompt. This presents a problem.
If neutron population in the core changed this fast, no human would be able to control it. Human reaction time is about .250 seconds. Thankfully, the other .1% of the neutrons in the core are delayed neutrons, born from the decay of fission products, not the fission event itself. These are technically any neutrons born after 10^-14 seconds. On average, it takes 12.7 seconds for the fission products to decay and produce a delayed neutron.
This population of delayed neutrons slows everything down. If you do the math to figure out the time it takes for a generation of neutrons to live and die, the delayed neutrons push that number to something manageable by humans, really close to the 12.7 seconds it takes to make them.
A prompt critical reactor is one that is creating new neutrons so quickly that the reactor would be critical even without the delayed neutrons. Unfortunately, those delayed neutrons do still exist, which means there are way too many neutrons in the reactor. This excess of neutrons causes the fission rate to exponentially rise. Basically, every last bit of fuel that can fission at that moment does. Power skyrockets.
This is what happened at SL-1 and what happened at Chernobyl, but for different reasons. I will probably write another post on Chernobyl at some point in time. For now, I am going to use the data from Chernobyl to explain this a little more, because more information is available for that event than SL-1.
The graph above shows several different reactor parameters at the time of the explosion at Chernobyl. The solid light blue line is reactor power. The y-axis scale is 0% at the bottom and 120% at the top. The x-axis is 1 second per interval.
So, in 2.5 seconds, power went from essentially 0% to 120% and was rising nearly vertically. Vertical lines on anything related to operating a reactor are rarely, if ever, a good thing.
Estimates are that Chernobyl maxed out at 10-100x it’s normal full power of 3000 Mwt before the pressure surge destroyed the reactor and stopped the chain reaction. The dashed blue line is another indication of power, delayed by a few seconds for some reason (I don’t know exactly how their instruments worked and where the Soviets got this data). Its scale is 0% to 50000%, so power goes from 0% to ~45000% in about a second.
SL-1 was a much smaller core and I don’t have any graphs for it, but they would look roughly similar. Design power was 3 MW. Estimates for power during the prompt criticality are 20000MW. The reaction at SL-1 stopped for the same reason Chernobyl’s did – all of the water instantly turned to steam causing pressure to rise to approximately 10000 psi, which destroyed the reactor.
TLDR version, critical reactors are at a steady power level. Sub-critical reactors are shutting down. Super-critical reactors are increasing power. Prompt critical reactors explode.
3 people died at SL-1. They were so irradiated from the explosion that they were buried in lead caskets. The bodies did not decompose like normal bodies would, because they had been almost entirely sterilized by the burst of radiation.
Modern civilian reactors in the US cannot go prompt critical. The control rod worth of any single rod cannot be so high that, if it were somehow ejected from the core, the entire reactor would go prompt critical. The fuel directly adjacent to that rod might, but not the entire core.
Some fuel damage would be expected, mainly cladding ruptures in the fuel the control rod was in (cladding is the outer wall of the rod the fuel pellets are in), but the core would remain intact. Safety systems installed at all US reactors would ensure core cooling was maintained following the event.
As always, this is all open source. Feel free to explore the links below.
Sleepy Neutron's Nuclear Knowledge 
Leave a comment