If there’s anything more amazing than the credulity of many reporters, it’s their pluperfect ignorance of all things technical, and their (and their editors’) indifference to getting technical matters straight.
Here’s a paragraph from the Time “Yellowcake Road” story (*):
The tale begins in the early 1980s, when Iraq made two purchases of uranium oxide from Niger totaling more than 300 tons. Known as “yellowcake,” uranium oxide is a partially refined ore that, when combined with fluorine and then converted into a gas, can eventually be used to create weapons-grade uranium.
There’s so much wrong with that paragraph it’s hard to know where to start.
Uranium occurs naturally in an ore called “pitchblende.” Pitchblende is mined, then refined into uranium oxide (U-three-O-eight), or “yellowcake.” That’s the end of the refining process. What happens next is called “enrichment.”
Uranium comes in several isotopes, of which two make up almost all of natural uranium: U-238, about 99.3%, and U-235, about 0.7%. The Canadian CAN-DU (heavy water) reactor can actually use the natural mix to produce power, but the light-water reactors the rest of the world uses need “enriched” uranium, with about 3% U-235 rather than the 0.7% that occurs naturally.
There are two enrichment technologies in current use. The one developed at Oak Ridge during WWII is called “gaseous diffusion”: it involves converting yellowcake into uranium hexafluoride gas, which is then run through a set of baffles. The slightly lighter molecules containing the U-235 move through the baffles slightly more quickly, so the gas that comes out of the baffles first is slightly enriched. Run it through again and again, and after a while the stuff is up to the 3% level that makes it “fissionable.” The uranium is then converted back to the oxide form and fabricated into fuel rods.
If you repeat the process many more times, eventually the uranium can be enriched to the 20%-U-235 level that makes it “fissile”: i.e., suitable for bomb-making. (That would be a pretty crude, and necessarily bulky, bomb: the stuff in a real, modern weapon is more than 90% U-235, or alternatively it’s plutonium. The stuff in a bomb is the pure metal, not the oxide.)
The alternative approach uses a centrifuge: here the heavier molecules move faster, but the separation principle is the same.
(I actually remembered most of this stuff from a previous life working for Les Aspin on the breeder reactor and related issues, but just to check the details I Googled “uranium enrichment, just as someone at Time should have done, and found this.)
So you can see that the paragraph from the Time story works passably as a mnemonic if you already know what’s happening, but that everything in it is slightly wrong and the whole infused with a kind of magical thinking. The conversion to uranium hexafluoride, which is mentioned, is an unimportant detail, while the enrichment, which isn’t, is the heart of the problem. Yellowcake is cheap (*); you can get enough to build a bomb for about $20,000. Separative work units (SWU’s, the measure of enrichment effort) are the expensive item.
It would have been much more informative, and much less misleading, to use the same number of words to say something like:
“Yellowcake — chemically, uranium oxide — is the end-product of the uranium-mining process. After costly and difficult enrichment by isotope separation, the uranium can fuel nuclear reactors or be made into bombs.”
The naive reader still wouldn’t know much about the technology — what’s “enrichment by isotope separation”? — but he would know what, for practical purposes, yellowcake is, and he wouldn’t be left with a vague image of bombs being made out of gas.
This sort of writing is the product of letting math-and-science-phobic English majors become reporters. It matters, as my friend David Kennedy keeps insisting, because only the habit of thinking concretely about technical matters allows you to get a feeling for what is, and what is not, technically possible. In the end, if you’re not a specialist, you’re going to wind up relying on the word of people who know more than you do, just as is the case when you’re trying to figure out a problem involving tort law or infectious disease or macroeconomics. A lot of this stuff is rocket science.
But it isn’t magic.