Every shop that builds something ends up with a scrap pile, and the scrap is almost never the good part. Sawdust, metal shavings, offcuts nobody wants. Oak Ridge National Laboratory has a scrap pile too, left over from making californium-252, and it’s probably the only scrap on Earth that gets reserved by customers before it even exists.
Californium-252 is the material that wakes up nuclear reactors. A freshly loaded core is so quiet that there’s almost nothing for its own neutron detectors to count, so crews put a californium source in to give the instruments something to read, and we’ve already covered how ORNL rolls the stuff into wire on a jeweler’s mill. ORNL’s the only source of it in the Western world, an industrial consortium seals it into finished sources, and demand keeps climbing.
This time I want to talk about what’s left in the tank afterward.
ORNL runs a californium production campaign every two years, and each one leaves behind three of the rarest substances anyone has ever bottled: picograms of fermium-257, micrograms of einsteinium-253 and -254, and whole milligrams of berkelium-249. The lab’s own comparison goes like this: a milligram is roughly a grain of sand, a microgram is a dust speck you’d need a microscope to spot, and a picogram weighs about as much as the DNA inside a single cell. ORNL only talks in those orders of magnitude, so I can’t tell you the exact amounts. Measuring absurdly small things is kind of this lab’s specialty anyway (it recently pressed a measuring tip into a poppy-seed-sized nuclear fuel particle one layer at a time).
According to ORNL, that’s the entire stock that ever goes on sale, and researchers claim it before it’s even been made, mostly through the Department of Energy’s Heavy Element Chemistry Program. Cristian Celis-Barros, the radiochemist leading the recovery effort, says the isotopes would otherwise count as waste. His job’s fishing those traces out of the leftovers without disturbing the californium everybody is actually paying for.
So who buys a picogram of anything?
Physicists, mostly. These isotopes sit at the heavy end of the periodic table, and if you want to discover an element nobody’s made yet, you pretty much have to start from one of them. ORNL’s berkelium-249 was the raw material behind the discovery of element 117, tennessine, and the lab says its radioisotopes also played a part in finding and confirming elements 114, 115, 116 and 118.
Fermium has had a weirder run. The element was discovered in 1952, ORNL measured it out decades ago, and then the scientific community more or less lost interest. Now physicists are asking for atoms again. What fermium’s actually good for, nobody knows yet. ORNL says it could turn out to have uses in medicine or industry, which is a polite way of admitting the same thing.
Some of this chemistry dated back to the 1950s
The age of the old methods surprised me. ORNL pulled these isotopes apart for decades with processes developed in the 1950s, and those stayed in service until a researcher named Miting Du replaced them about five years ago. Du retired in 2025. Celis-Barros has been building on that work since, and his latest tweaks landed during the most recent californium campaign.
Most of the upgrades run through chromatography columns, and the idea’s simpler than the name suggests. You pour a solution through a tube packed with resin that grabs certain elements and lets everything else wash past, so the thing you want stays trapped while the junk rinses away. Swap the resin or tweak the conditions, and the column catches something different.
The berkelium now comes out green
For berkelium-249, Celis-Barros added one extra column run to Du’s process. The new pass holds onto the berkelium while leftover impurities wash through. The goal was chemical purity on top of the radiochemical purity the old process already delivered, and you can see the payoff from across the room. The solution used to come out a murky yellow. Now it’s a solid green crystal.
Celis-Barros borrowed the idea from his earlier work purifying promethium-147, another ORNL product, and the lab says the revised berkelium step is still preliminary, with more testing ahead before it gets adopted for good. Judging by the photos ORNL released, I’d say the before-and-after sells itself.
The einsteinium went through the same type of extra column, with the conditions tuned to split it away from the californium. By the end of the campaign, 95 percent of the einsteinium was coming back on every column run, which ORNL says is significantly more than past campaigns managed.
Fermium took a different route. The resin radiochemists usually reach for in this kind of extraction is DGA, short for diglycolamide, and for fermium-257, Celis-Barros swapped it for a different resin entirely. So what did he swap in? ORNL hasn’t said, which I’d love to know more about, but the result’s hard to argue with: a chemical purity of about 99.93 percent.
And my favorite detail in this whole story is how he knew it had worked. He couldn’t see anything. At picogram scale, a sample you can spot with the naked eye means contamination, so a vial that looks completely empty is exactly what success looks like.
The whole job happens in gloveboxes
You might assume this work happens behind the thick walls of a shielded hot cell, like the bulk californium separation does. It doesn’t. The byproduct recovery runs in gloveboxes, which cost far less to operate, and the trade-off shows up in the schedule. The solutions still hold micrograms of highly radioactive californium, so the chemists have to move fast, because every hour of handling counts against their maximum allowable radiation dose. Once a worker hits that limit, they’re done, finished or not.
That’s really what the whole project is for. Celis-Barros says the team is simply trying to “come up with processes that are more efficient and expose workers to less radiation,” and the American Nuclear Society’s Nuclear Newswire reports the improved separation has physicists more interested than ever in getting einsteinium and fermium split from each other.
Oak Ridge gets another shot soon enough. The next californium campaign’s set for 2027, and Celis-Barros says the team now knows how to prepare and plans to recover more than it did this time around. ORNL published the current results on September 22, 2026.
