A new safeguards approach for enrichment plants

How are enrichment plants safeguarded today?
Quite possibly, uranium enrichment plants represent one of the greatest challenges to the non-proliferation regime. If the site is large enough, a state could amass the necessary nuclear material for one weapon in a matter of days, not months or years. The non-proliferation norm which has held its ground since 1968 will be seriously eroded the more states that possess this break-out capability. So clearly, most thought has gone into the problem of how to verify the peaceful use of this dangerous technology? Sadly, the answer to this question is a simple "no".

Enrichment plants are safeguarded using a methodology that was developed by a small group of states in the 1980s under the rubric "The Hexapartite Safeguards Project". This group was like-minded and allied, which is clearly reflected in the procedures they agreed.

Presently, there are about thirteen inspections per year at a typical uranium enrichment facility. Once a year, the physical inventory is audited by IAEA inspectors through a so-called "physical inventory verification". In addition, monthly routine inspections are conducted to verify material flow. These inspections, called LFUAs have an element of unpredictability.

Anyone interested in learning more about the Hexapartite Safeguards Project should read the highly informative "On-Site Inspections: Experiences from Nuclear Safeguarding" by Wolfgang Fischer and Gotthard Stein.

Material balances at a typical facility
At a small facility with 3,000 centrifuges capable of 2 SWU/year, the material flow could look something like this. Throughout the year, 15,75 tons of uranium hexafluoride gas will be received by the facility. This material will be processed, resulting in a product of some 1,95 tons of gas enriched to 3,6 per cent U-235, and some 13,80 tons left in the tails (depleted to about 0,3 per cent U-235). The highly corrosive gas is transported and handled in special steel cylinders. In the United States, the standard 48" cylinder holds about 10 tons of gas, and customer cylinders about 2,5 tons. If this smaller enrichment plant uses 2,5 ton cylinders, the feed will be contained in about 7 cylinders, the enriched product in one cylinder and the tails in six cylinders. They will all be weighted when they arrive at the facility and before they leave the facility. The idea is that if you know how much material is coming in, and how much is going out, you'll know how much is left inside the facility.

Now also assume the presence of an undeclared plant with some 1,000 centrifuges capable of five kilograms of separative work per machine and year. This plant would be small, requiring only some 1,000 square meters of floor space, and less than 500k kWh of electricity, which is significantly less than the electricity used by a medium sized manufacturing plant. This plant is conserving the uranium loss by working with a lower tails assay (0,15 per cent). It would need about half the output of the first plant, or one ton of enriched uranium per year, to produce some 25 kilograms of weapons-grade uranium. The supply can be produced in a single transport cask, or in two or three 500 litre drums. Despite investing more separative work into the product, it would be able to enrich the uranium in a little more than five months.

Some readers will now ask: but clearly the International Atomic Energy Agency would find out that one ton of uranium hexafluoride gas is missing from the first facility? Perhaps so, if the deviation is as large as half the throughput of the facility, but probably not if the deviation is smaller. It is fairly simple for the facility staff of the first plant to work around the safeguards. They will simply declare delivery receipts that show that the facility only received seven and a half tons of hexafluoride gas under the year. If the facility operator claims that he or she received half the material, then the product should only be one ton, not two. Under present safeguards arrangements, the Agency can only determine that the enrichment level is as declared and that the product matches supply. There is no reliable way to measure the flow of materials through the enrichment plant.

Bad news comes with good news
In other words, it is fairly easy for the facility operator to pass surplus - in essence undeclared - material through their cascades. The inspectors would have a very low chance of detecting this, as long as the enrichment levels to not deviate from declared levels, and the operator could account for it by understating his receipts. This material could then be shipped off to another, clandestine, facility for HEU production. There are few immediate fixes to the problem. One is to station a resident inspector at the facility, who can verify flow changes in real time. Another solution is to install electronic mailboxes at the facility, which sends off real-time process inventory listings to the IAEA in Vienna. A mailbox system is shorthand for a computer network or other arrangement in which operators provide operational information in a form that cannot be changed. This information would include the cylinder inventory as well as information on facility status. It is not impossible to beat such a system, but it makes it more difficult. The operator would need to plan a diversion more diligently.

The good news is that it is difficult to use a facility for the production of undeclared highly-enriched uranium. It can be done through reconfiguring the cascades or to operate the cascade in recycle mode. The most obvious diversion scenario is not to declare product or feed material. However, if declared material is used for feed, it can be listed as material unaccounted for, as an shipper-receiver mismatch, or as a operator-inspector difference statistic. To put it in a simpler way, it is relatively easy to tamper with the books. The Agency has proposed a range of measures to counter this scenario, the most famous being to install some form of containment and surveillance inside the cascade hall (cameras, cameras, cameras). This has been proposed in Natanz, but the Iranians have not warmed to the proposal.

Another piece of good news is that the IAEA does not see small scale research and development enrichment facilities as a major problem from a safeguards point of view, since the inventory is small and operations are discontinuous.

What can be done?
The prevailing view amongst safeguards professionals today seem to be that randomly timed inspections may help. However, these inspections must be supplemented with unattended and remotely monitored instrumentation which will provide a continuous detection capability.

Cameras are very important. This is primarily because inspectors are not allowed to see everything that matters. Cascade piping arrangements are not as transparent as was assumed during the early 1980s. For instance, trace lines assigned to sampling and chilled water systems might be used to continually withdraw cascade material, but they are not surveyed during an LFUA.

At present, there are no forms of flow measurement in the cascade area of the enrichment plant. This does not mean that they should not be used. An in-line mass flow meter could provide trend information related to cylinder filling or emptying outside the cascade hall. Thermal mass flow detectors are accurate. They monitor the current needed to maintain an element at constant temperature in the process stream. A change in mass flow will lead to a corresponding change in current.

Work on a new safeguards approach for enrichment plants have started in earnest only in the past two years. Such an approach is urgently needed. Emerging nuclear fuel suppliers, such as Iran, should seriously consider implementing new approaches to safeguarding their facilities, than clinging on to an old concept, which is not as suitable to provide reasonable assurances of non-diversion of nuclear material.