A number of you have requested posts on JCPOA verification and the “24-day” issue. A way to start is with Jeffrey Lewis’s request for how environmental remediation relates to JCPOA verification. It’s something that I will need to refer back to in discussing those issues. And it’s clearly something that numerous commentators have no idea about. Basically, the requirements for sampling should be pretty much the same for IAEA inspections as for environmental remediation. Both have to stand up to legal scrutiny.
I’ll use three sites as examples: a metal plating bath outflow that was one of my responsibilities at Los Alamos, the Parchin site in Iran, and Al Kibar. I’m not making any big points here about Parchin and Al Kibar. I am using them to show what sampling requires.
Sampling is easy, right? You dig up some soil and put it in a baggie, or you swipe a wall with a tissue, and then you send it to the analytical lab and they tell you. BZZZZT! WRONG!
Sampling starts at a desk. First, you have to figure out the question you are trying to answer. The environmental remediation questions are pretty standard – what is there, how much, and where it is spread to – but the IAEA’s questions tend to be more varied. At Al Kibar, the question is whether there was a reactor there before the Israeli raid and the Syrian cleanup of the site. The situation at Parchin is more complicated. Three types of experiments are alleged to have been done in a containment chamber inside a building, after which the Iranians made many modifications to the site, including modifications to the suspect building, soil removal, and asphalt overlay. The basic question is which, if any, of those experiments took place there.
Second, you have to figure out what kind of samples you need to answer the question. For the plating outflow, that meant going to the archives to find out what kinds of metals and other chemicals were involved in the plating operation, what was released in the outflow, when and for how long. You also need to know what kind of samples the analytical laboratory will need to get good analyses. If you spend days getting 10-gram samples and the lab needs 100 grams for the analysis you want, well, you’ll have to do it again. And the IAEA doesn’t always get to do it again.
At Al Kibar, the materials of interest, which would be present only in very small particle form after explosions and earthmoving, if at all, would be materials found in reactors. Uranium would be the most indicative, and enriched uranium even more so. Other reactor construction materials like graphite, steel, and zirconium would also be included in the plans for analysis. Soil was all that was left after the attack and cleanup, and there are many ways to take soil samples. The overhead photos and any information available on the mode of bombing and the munitions used would be helpful in determining where to take the samples. Were there indications of material being dispersed or bulldozed toward particular locations? Is the material likely to be in surface soil? Deeper?
At Parchin, each scenario – initiator testing, hydrodynamic experiments, or spherical implosion testing – would result in different materials dispersed. For initiator testing, polonium would be pretty much gone after more than a decade, with a half-life of 183 days, but beryllium would remain. For hydrodynamic experiments, uranium or plutonium, possibly other heavy metals like tungsten. Spherical implosion testing might not contain anything particularly diagnostic, although the explosives might leave a telltale residue. The containment vessel and its building would retain most of the materials and would be the most desirable sampling areas, but if there was a containment vessel, it seems to have been removed. Swipe samples would be taken from the vessel and building walls would be taken. If the walls were painted over, paint chips might contain the materials.
In cleaning up the building, the Iranians may have concentrated the residues in outside soil when they washed out equipment, as seen in overhead photos. That area was later asphalted over, which should help to keep the residues in place. So sampling would be in that area.
Any other information, such as records, would be invaluable. But if the Syrians made full records available on Al Kibar or the Iranians on Parchin, sampling would probably not be necessary.
We had statisticians to help us design our sampling plans. They calculated sampling patterns that would help to delineate how much soil needed to be removed. The plating outflow was a small stream that trickled down the porous tuff that underlies the soil of the Pajarito Plateau. Its track, about a foot across, was stained red from the metals. We sampled the track to see what metals were there and also outside the track to determine whether metals had migrated there.
The questions for Al Kibar and Parchin require less information about how far the residues have spread, but statisticians could calculate which areas would be best to sample and how deep they should be sampled.
Standard procedures exist for various types of soil sampling. You might want a sample that is mostly from the surface or mixes together the soil from, say, the top six inches. Or a profile of where contaminants are in the soil column. There are procedures and equipment for that, ranging from trowels, stainless-steel bowls (top photo), and coring augers up to truck-mounted soil punches. For both Parchin and Al Kibar, surface samples might be best for undisturbed areas and samples at several depths where soil has been piled up or in that runoff area at Parchin.
It’s all written down. The numbers and types of samples are listed in a sampling plan and checked by the sampling and analytical teams to make sure they can do it. You can’t take a four-foot-deep soil sample if there is solid rock at that depth. Next-best locations may be specified in case some locations can’t be sampled. The statisticians need a certain number of samples to do their job.
Third, the sampling team collects the samples. They have packed all the equipment they will need before going into the field. That includes appropriate containers for the materials, often plastic bags or boxes, specialized containers for cores if they are to be preserved whole; computer-generated labels for the samples, corresponding to the sampling plan and any codes the analytical laboratory needs; sealing materials, writing materials as necessary, materials to clean sampling implements as necessary, and a computer to log in the samples. They may have paper copies of chain of custody forms.
The team uses surveying methods to accurately locate the sampling points. Each sample is sealed in a container so that it cannot be contaminated. It is labeled, and more than one person will check the sealing and labeling to make sure they are done properly. They will be packed in carrying containers in specified ways.
If the results of sampling and analysis are to support a legal argument, the samples must have a chain of custody, similar to what is used for evidence in criminal law. A designated custodian on the sampling team signs the paperwork, although I suspect this is all done by computer now. Dates and times of transfer to another party are specified, for example from the sampling team to the analytical lab.
Fourth, the samples are analyzed. There are standard procedures for this, too, and many machines that can measure amounts down to almost single atoms or molecules. In the three examples, metals are most important, and, in many ways, the easiest to analyze for.
The samples will probably be concentrated, by separating the soil into different particle size fractions or by dissolving parts of the sample and leaving others. The final step will most likely be mass spectrometry, counting of atoms of different metals and even isotopes. Isotopic analysis wasn’t necessary for our plating bath outflow, but it was for the uranium particles reported by the IAEA at Al Kibar, because the IAEA would have wanted to know if that uranium was natural, enriched, or depleted. It might be useful at Parchin, too.
The statisticians then work over the results. This was an essential step for environmental remediation and resulted in maps of where the contaminants were. Whether it is for Parchin and Al Kibar depends on how the sampling plans were designed. There may be further checks of the sampling, analytical data, and statistical treatments.
At Al Kibar, the IAEA hasn’t said what size the uranium particles they analyzed are, but I suspect that they were found in soil samples by electron-microscopic examination combined with detection of various metals, that they were of the same size as the soil grains or smaller. The IAEA hasn’t said if an isotopic analysis was done, but that is both likely and possible.
Finally, I know I said I wouldn’t do this, but as I’ve written this up, two substantive points occur to me on Al Kibar and Parchin. They illustrate why you need protocols for sampling.
Because of these protocols, Peter Jenkins’s informant’s story is highly improbable. The swipe sample that he claims is the source of the Al Kibar uranium particles couldn’t have been substituted for soil samples because there would have been a sampling plan and chain of custody. You don’t just say, “Oh, here’s the sample you should analyze.” It has to be a sample according to the plan. Might a sampling team pick up something additional if it’s interesting? Sure, but notes would be taken to describe how it was collected, and it wouldn’t be substituted for a sample in the plan.
At both Al Kibar and Parchin, there has been a great deal of site modification. That means that only positive findings, like the uranium at Al Kibar, can be meaningful. If nothing is found, that could mean a clean bill of health, but it could also mean that the materials were indeed removed by the cleanup. So even if Parchin is sampled and no telltale metals are found, we can never be sure that the alleged experiments were not done.