by CKR
I’m seeing some misconceptions out there, like this one
If I understood Professor von Hippel correctly, then, 3,000 centrifuges running in parallel could enrich enough uranium for one bomb in four months, or three bombs in a year.Just to show I’m not picking on Dave, I’ll also note that I saw a comment somewhere else about running one or two centrifuges while you build up the rest. Neither of these quotes is quite wrong, but they’re not quite right either.
Most of the explanations of gas centrifuges on the web concentrate on a single centrifuge. But the entire installation and how things are hooked up is an important part of the process too, and that gets less attention.
In order to do industrial-scale isotope separation, you have to hook up many centrifuges into cascades. Cascades are not easy to explain, but I’m going to try.
The mass difference between uranium-238 hexafluoride molecules (mass 352) and uranium-235 hexafluoride molecules (mass 349) is 0.85%, very small. That means that a single centrifuge can’t separate all the U-238 from all the U-235. It can produce two streams, one slightly enriched in U-235 and one slightly depleted. I’ll call the first hexafluoride+ and the second hexafluoride-. Each of those streams goes into another centrifuge, and you get hexafluoride++, hexafluoride+-, hexafluoride-+, and hexafluoride--.
You might think, “and so on,” but simply putting two outputs into two more centrifuges loses about half the U-235. That consigns the first hexafluoride- stream, which contains almost half the U-235, to irrelevance.
Depending on how much separation each centrifuge can do, those +- and -+ streams from the second centrifuges may be the same or they may be different. They may be able to be combined, or with earlier streams. The tubes on the centrifuges in the photo (stolen from David Albright’s ISIS) give you some idea of the complexity. The photo is of one cascade, made up of several centrifuges. Each cascade can be regarded in the same way as an individual centrifuge—input gives an enriched stream and a depleted stream, and then they are hooked up similarly.
If you hook them up wrong, you’re throwing away the energy that has gone into the previous separation. And then there’s the practical problem of getting all those fittings sealed properly. It’s much trickier than sealing the oil filter on your car. Any leak allows atmospheric moisture in, which will react with the uranium hexafluoride and deposit solids in those skinny tubes, plugging them up, or on the centrifuge rotors, tearing them apart, along with HF, which will corrode more seals…
And that’s how you do it.