What do we really know about Iran’s capability to reconvert triuranium octoxide (U3O8) enriched to 20% U-235 back into UF6 feedstock that can be further enriched to produce weapon-grade uranium? Can Iran do it? And if so, how fast?

The answer matters considerably, as Iran, Israel, and the P5+1 will make decisions this year, based in part on their assessment of risk, about the fate of current efforts to negotiate a comprehensive crisis settlement.

Mark Hibbs
Hibbs is a Germany-based nonresident senior fellow in Carnegie’s Nuclear Policy Program. His areas of expertise are nuclear verification and safeguards, multilateral nuclear trade policy, international nuclear cooperation, and nonproliferation arrangements.
More >

In the policy world, there are two opposing views being expressed, whether they are informed by the facts on the ground in Iran–or not.

Advocates of stepped-up diplomacy with Iran argue that Iran, by not accumulating 20%-enriched EUP from the Fordo enrichment plant as UF6 but instead converting some of it to U3O8, is signaling to the powers its willingness to compromise and de-escalate the crisis. In U3O8 form, they argue, the material would be less directly usable should Iran want to dash to a bomb, because Iran would have difficulty reconverting the oxide to UF6, especially if the oxide had been fabricated into finished research-reactor fuel.

Iran’s determined adversaries assert to the contrary that there is no nonproliferation benefit in Iran converting its 20%-enriched Fordo output to U3O8 because Iran could reconvert the material back to UF6 easily and in a hurry.

Iran has described its converting of the UF6 into U3O8 as a confidence-building measure.

U3O8 and Israel’s ‘Red Line’

This week, a former Israeli intelligence official claimed that Iran could within a few days reconvert its U3O8 back to UF6, implying that Iran has already crossed a “red line” set by Israeli Prime Minister Benjamin Netanyahu–the production of enough 20%-enriched uranium which could be enriched to weapon-grade and fashioned into a nuclear weapon. Three days later, in a Reuters interview alerting the outside world it is prepared to continue talking with the P5+1, an Iranian negotiator made known that Iran intends to continue converting the Fordo product to U3O8 for use as fuel for five research reactors. These reactors presumably would include the existing Tehran Research Reactor (TRR) plus four more reactors that, according to Iranian press reports  in 2012, Iran’s President Mahmoud Ahmedinejad had ordered to be built. Whether or not Iran ever builds those four reactors, they provide Iran a rationale to continue enriching uranium to 20% and converting the EUP to U3O8. There is also the research reactor at Arak, but it will not require 20%-enriched uranium to operate when it is finished.

As of late February, the last time when the IAEA reported to the Board of Governors on safeguards implementation in Iran, the status of Iran’s inventory of 20%-enriched uranium produced at Fordo was this, as I described in a post earlier this month:

After commencing with the enrichment of uranium to 20% U-235 in early 2010, Iran accumulated about 150 kilograms of EUP at this enrichment by the end of 2011, and it crossed the 200 kg threshold sometime in the middle of 2012.  The latest report from the IAEA in February says that Iran had produced 280 kg of UF6 enriched to 20% U-235, of which 167  kg was still in the form of UF6. Virtually all of the rest has been introduced into the reconversion plant to produce U3O8 for fuel fabrication. When the IAEA accounted for Iran’s declared activities in February, the plant had produced U3O8 containing 50 kg of uranium, leaving about 60 kg of uranium in the process inventory. According to the IAEA data, the current rate of production of feedstock at this enrichment level in its centrifuges is about 15 kg per month.  

By late February, then, Iran had processed about 110 kg of its accumulated UF6 inventory enriched to 20% into U3O8 and other intermediary chemical states. At current rates of production at the Fordo enrichment plant and chemical processing facilities at Esfahan, by the end of this year Iran might accumulate a U3O8 inventory containing approximately 175 kg U enriched to 20% U-235–but only if it is assumed that all the uranium fed into the conversion line would be converted to U3O8. Data from the February IAEA report suggest that the real conversion factor from UF6 to U3O8 is far less.

According to Susan Voss based on data in IAEA reports on Iran’s safeguards implementation Iran has lost 61% of its uranium in feed material during the conversion of UF6 to U3O8 for the TRR.

Olli Heinonen however believes Iran’s present conversion factor is higher.

The IAEA reports provide the amount of UF6 moved (and “released”, which means the cutting of the IAEA seals from the UF6 cylinders) to [Iran's Fuel Plate Fabrication Plant] but it does not give a full breakdown for material in each step of the process. The reports give the amount of U3O8 produced at certain point of time, which is just one part of the material balance equation. There are all the reasons to believe that the Iranian engineers, with two decades of experience on uranium conversion, can achieve a better yield than 39 %. In addition, 20 % enriched uranium is valuable material. Like the other fuel producers, [Iran has] a small process, at least on the drawing board, to recover uranium from the wastes (albeit only a few per cent of material should end up in wastes).

In a statement before the Institute for National Security Studies in Tel Aviv on April 22, Amos Yadlin, former head of Israel’s Military Intelligence  Directorate, asserted that Iran in “less than a week” could convert its 20%-enriched U3O8 into bomb-grade “nuclear material” for a weapon.

In separate comments made to Israeli radio, Yadlin appeared to suggest that right now 80 kg had been processed into U3O8 and was therefore available to be reconverted to UF6.

Iran’s Capabilities and Options

We may assume that Yadlin’s remarks in some quarters will be interpreted to drive a stake into the heart of any forthcoming compromise deal with Iran, challenging those who argue that Iran has demonstrated self-restraint in not stockpiling 20%-enriched UF6. Iran’s capabilities for reconverting the 20%-enriched U3O8 back into UF6 feedstock for nuclear weapons fuel therefore need to be understood and the following questions need to be answered: 

Does Iran now have the means to reconvert its U3O8 to UF6?

Yes. Do not be confused by the terms conversion and reconversion. I have heard it said: “Iran can convert the Fordo enriched uranium from UF6 to oxide but it cannot reconvert the oxide back to UF6.” Not true. The chemical processes corresponding to “conversion” and “reconversion” are more or less identical. Iran has lots of experience converting its uranium ore concentrates to UF6, and Iran can likewise convert U3O8 obtained from UF6 back into UF6. In both cases, the feedstock for conversion is U3O8. In the first instance, it is milled from natural uranium ore. In the second, the feedstock is oxide that has been converted back from UF6 which was previously enriched.

How would Iran process its U3O8 into UF6?

Iran would have a number of options, but there are two basic ones.  They principally differ in how to convert U3O8 to the intermediate product UO2. To convert the UO2 to UF6, the process would be the same for both options.

One option would be to reconvert U3O8 into UO2 using a process similar or identical to that used at Iran’s Uranium Conversion Facility (UCF) at Esfahan, which Iran has operated to produce its UF6 feedstock for centrifuge plants at Natanz and Fordo. The U3O8 would be dissolved in nitric acid, producing an aqueous solution of uranyl nitrate hexahydrate [UO2(NO3)2 . 6H2O]. In some versions, this would then be mixed with tributyl phosphate to remove the uranium in the form of uranyl nitrate. The nitrate can be converted to UO3 either by evaporation or treatment with ammonia. The UO3 is in turn converted to UO2 in fluidized reactors by reduction with ammonia gas at high temperatures.

A second option would be a dry process to expose the U3O8 to very high temperatures in the presence of hydrogen gas. The endothermic reaction of U3O8 with H2 would result in UO2 and water. It is likely that Iran has studied and may have mastered this kind of process at laboratory scale. If Iran has mastered it, less time may be required to reconvert the U3O8 than by using a wet process, because fewer steps would be needed. Iran might favor a dry process route because its enriched U3O8 contains few or no impurities, obviating the need to do solvent extraction. The impurities, such as oxidizing metals, would have been already removed at UCF prior to enrichment of the NATU at Fordo.

For both wet and dry options, after the material is converted to UO2, it would be reacted with anhydrous hydrogen fluoride (AHF) to produce UF4, and the UF4 would in turn be fluorinated to result in UF6.

Iran has investigated several process chemical options for doing this since the 1980s. Most of the processes have been applied elsewhere in the world, and in all the nuclear weapon states, beginning in the 1940s, and they are well-known.  In addition to experience gained at UCF since the mid-2000s, Iran a decade before operated a small chemical conversion lab to produce UF6, and Iranian scientists have also worked on uranium conversion chemistry in its so-called “Green Salt Project.”

Where would Iran do the U3O8-UF6 reconversion?

In theory, Iran could use its existing and declared conversion infrastructure at Esfahan to convert the U3O8 beginning with nitric acid dissolution and ending with production of UF6 gas. In practice, because any re-conversion in a safeguarded facility of a discrete inventory of previously enriched U3O8 would, if detected or declared, prompt IAEA inspectors’ concerns (the UCF is routinely monitored), reconversion would more likely take place in an undeclared facility dedicated to process enriched U3O8.

Should Iran choose to reconvert the U3O8, it would have other motivations to do it in a dedicated, small installation. The geometries of such a facility could be designed to minimize the risk of a criticality accident, which could occur during the processing of 20%-enriched feedstock. At a bulk-handling facility such as UCF, designed to process NATU, criticality management would be more challenging, and the risk of an accident, especially if enriched uranium were converted under duress, would be greater. The kind of issues Iran would face are illustrated here.

A small facility would be best to permit batch processing, allowing personnel to most effectively control off-gas and ventilation systems needed to cope with volatile hydrogen and fluorine gases involved in the conversion of UO2 to UF6. Iran might use a fairly simple process, similar to that used at the JCO fuel processing plant in Japan, which in 1999 suffered a criticality accident when buckets filled with uranium solutions were carelessly handled by personnel.

How much time would Iran require to reconvert the enriched U3O8 to UF6?

Yadlin, cited as having told the Institute for National Security Studies that in Iran the reconversion of U3O8 to produce bomb fuel could be “completed in less than a week,” walked back this estimate in a subsequent radio interview to “between one and two weeks.” That’s more realistic. Experience from the uranium conversion industry and R&D sector outside Iran would suggest that Iran might be able to convert about 100 kg of U3O8 to UF6 in about two weeks–provided, however, that the work was carried out in a small facility using a dry process without purification, whereby perhaps three batches would be consecutively processed. Use of other processes and a larger installation might lengthen the time required to reconvert the material. Regardless of whether Iran would select a wet or dry process, the most time-critical process step would likely be the production of UF4 from UO2 because of the comparatively slow reaction time for AHF and UO2.

The IAEA Safeguards Glossary includes a table for estimated material conversion times to produce finished nuclear weapon metal fuel components using various fissile material feedstocks. Conversion time is defined as:

the time required to convert different forms of nuclear material to the metallic components of a nuclear explosive device. Conversion time does not include the time required to transport diverted material to the conversion facility or to assemble the device, or any subsequent period. The diversion activity is assumed to be part of a planned sequence of actions chosen to give a high probability of success in manufacturing one or more nuclear explosive devices with minimal risk of discovery until at least one such device is manufactured.

For “U-233 oxide and other pure U compounds,” the conversion time is given by the IAEA as “order of weeks (1-3).” A footnote specifies that for pure compounds it would be closer to one week and for “mixtures and scrap” material it would be closer to 3 weeks. This implies that to avoid detection a determined proliferator should take no more than three weeks to process pure uranium oxide into finished metal components.

If Iran processes its 20%-enriched U3O8 into fuel for the TRR–which Iran says it is doing and IAEA reports have verified–wouldn’t that serve as an effective proliferation barrier?

If Iran uses 20%-enriched U3O8 to make fuel for the TRR, converting that fuel material back to UF6 might require just a few more days than is needed to reconvert U3O8 powder.

The fuel for TRR consists of aluminum plates containing U3O8-Al fuel in a matrix. Iran would probably make this fuel using hot presses to bind the fuel material and the aluminum. To get at the U3O8 after fuel is fabricated but not yet irradiated, Iran could dissolve the fuel in a stainless steel vessel containing a caustic solution like sodium hydroxide, filter the mixture, and then dry and recover the oxide. Iran would have to be careful in handling the large amount of hydrogen gas that would be generated by dissolving the fuel.

Were the U3O8 to be fabricated instead into ceramic fuel using a sintering process for other reactor types, that fuel would be more difficult to break down and dissolve and it would potentially take longer to react with fluorine. Were Iran at some point in the future to make a deal with foreign governments and fuel suppliers including the supply TRR-type fuel, the proliferation barrier against diversion of enriched fresh fuel would be strengthened if silicide fuel were required instead of U3O8-Al fuel, because the silicide would be more difficult to dissolve.

The Bottom Line

Iran could process its entire inventory of 20%-enriched U3O8 to produce UF6 in a matter of a few weeks, the fruit of Iran's cumulative nuclear chemistry R&D and industrial-scale experience over three decades. There are some uncertainties about how great Iran's production losses would be, should it decide to reconvert the material.

An inventory of 20%-enriched uranium in Iran consisting of U3O8 reconverted from output from the safefguarded Fordo enrichment plant would be under IAEA safeguards. If it were associated with the UCF, it would be subject to physical inventory verification (PIV); if collected in small containers, it would likely be put under seal. The conversion of a portion of Iran’s U3O8 inventory into UF6 and subsequent re-enrichment could in theory be built into any of a number of break-out scenarios. How Iran would in fact behave can only be a matter of conjecture. Were Iran to inform the IAEA it intended to remove seals or reconvert the material into UF6, that step would immediately precipitate a crisis. If Iran were ever to decide to divert safeguarded enriched uranium to make a nuclear explosive device,  in addition to the risks of detection which would pertain to that action,  it would have to consider whether there would be any advantage in reconverting any of the U3O8 to UF6.

This article was originally published in Arms Control Wonk.