Last time I promised you a machine that gets harder to start the longer it has been off.
On the night of 26 September 1944, at Hanford in eastern Washington, the world’s first full-scale plutonium reactor went critical, ran up toward power, and then over the following hours faded and stopped.
“The reactor went dead, just plain dead,” Leona Woods Marshall remembered. “Everybody stood around and stared.”
Nothing was broken. And some hours later, with nobody touching anything, it started coming back — and once running, began to die again, on the same schedule.
A pulse nobody designed
Enrico Fermi’s group worked through the night on it. John Wheeler had warned in advance that the reactor might make its own enemy: fission manufactures several hundred new isotopes, and if any of them were hungry for neutrons, the reaction would be eating its own supply.
One of them is. Xenon-135 has a capture cross-section of about 2.6 million barns — the largest appetite for slow neutrons of anything known. In the trade it is not a contaminant but a poison, which is the right word: a small quantity, in the wrong place, stops the machine.
That much is the famous story. It is not the interesting part.
What is actually holding still
While a reactor runs at steady power, the xenon in its fuel sits at a constant level. That sounds like a stock: a quantity accumulated and sitting there. It isn’t. It is a race between two rates that happen to be tied.
On the production side, almost none of it comes from fission directly — only about a quarter of a per cent. Roughly ninety-five per cent arrives second-hand, as the decay product of iodine-135, which is made in quantity, and which takes its time: a half-life of six and a half hours.
On the removal side there are also two routes, and this is the sentence the rest hangs from. Xenon-135 decays on its own, with a half-life of about nine hours. But it is also eaten — such a voracious absorber that at operating power the neutron flux destroys it faster than its own decay does. The reaction it is poisoning is the main thing removing it.
So the steady level is not a reservoir. It is a standing wave in a stream of atoms: made from iodine at one rate, burned and decayed away at another, holding a constant height while not one atom stays.
Then you press the button
Now shut it down, and look at what you have actually switched off.
You have stopped the quarter of a per cent that came from fission. The real source is the iodine already sitting in the fuel, and iodine-135 does not care whether the reactor is running; it goes on decaying into xenon on its own six-and-a-half-hour clock, with no reference to anything.
What you have switched off is the removal. The flux is gone, the burning stops, and the only drain left is xenon’s slower decay. You have taken away the outflow and left the inflow running — so the poison does not begin to clear when you stop the reactor. It begins to rise.
It rises for about eleven hours — the figure falls straight out of the two decay constants — and only when the iodine has drained enough that xenon is dying faster than it is born does the level turn over. Twenty hours after shutdown a reactor is back to the burden it carried while running. Three days out, it is effectively clean.
The size of the hole
Here is why a control room cannot simply pull rods and climb out.
No. 116 introduced the operator’s unit of reactivity, the dollar: the entire margin between a reactor you can steer on a human clock and one that runs away in milliseconds. In a pressurised-water reactor a dollar is about six hundred pcm.
The xenon peak after a full shutdown can add as much as twenty-five hundred pcm of negative reactivity. About four dollars — several times the whole span of ordinary control, arriving over eleven hours, from a gas that was not there when you stopped.
If the reactor has excess reactivity in hand, it climbs out. Late in a fuel cycle it often does not, and then there is nothing to do. The condition has a name, xenon precluded startup, and so does the window: xenon dead time. You wait it out.
The reactors that never fall in
The pit is a disease of powerful reactors. Below a thermal flux of roughly 5 × 10¹³ neutrons per square centimetre per second, burnup was never doing much of the removing in the first place — decay was — so switching off takes away nothing much, and there is no peak at all. Weak reactors stop and start whenever they like.
Which is why power reactors are commonly held near that flux. Part of what limits how hard the most concentrated energy source we have is allowed to run is the requirement that it still be able to start on Tuesday.
≈
One loop I’m watching
Next: a fluid that is only liquid while you keep hitting it, and turns solid the instant you stop — a material whose entire mechanical identity is a rate, not a state. Next time.