Iran’s Uranium: Don’t Panic Yet.

By Ivan Oelrich and Ivanka Barzashka

Last week, the New York Times and the Financial Times USA ran stories that implied that Iran had been hiding enriched uranium and had been caught red-handed during the most recent International Atomic Energy Agency (IAEA) physical inventory inspection. While supposedly based on the IAEA report (GOV/2009/8), the articles more closely followed the ISIS analysis of the report.  [Jeffery Lewis, as usual, also has good analysis and comments on Arms Control Wonk.] The IAEA report itself raises few alarm bells. Yes, the Iranians are continuing to enrich uranium; yes, they claim it is exclusively for a civilian nuclear reactor program, a claim for which no one can provide credible assurances, and, yes, every day they enrich uranium, they are closer to having enough for nuclear weapon capability, once that political decision is made. But the IAEA report does not reveal any sudden jump in enrichment capability or even uranium inventory and it goes out of its way to say that the result of the inspection is consistent with what was previously declared by Iran, within “the measurement uncertainties normally associated with enrichment plants of similar throughput”. So what is the issue here?

The Financial Times headline, “Iran holds enough uranium for bomb” with the subtitles “UN report reveals leap in nuclear stocks” and “Capacity breaches Israel’s ‘red line’ limit,” and the New York Times headline, “Iran Has More Enriched Uranium than Thought” are both more provocative than warranted by the IAEA report itself. Both articles report that, in the most recent IAEA report from 19 February, the estimated inventory of low-enriched uranium (LEU) had jumped by a third. The New York Times said that the IAEA had “discovered” an additional 460 pounds (or 209 kg) of LEU. This number is wrong to begin with because the IAEA reported an additional 209 kg, not of uranium, but of uranium hexafluoride (UF6). UF6 is about 68 percent uranium, so there is only an additional 142 kg of uranium in 209 kg of UF6. We will come back to this.

Reasoning Behind the Headlines

“Discovered” is stretching. The origin of that 209 kg is the difference between the amount of enriched UF6 reported in IAEA documents GOV/2009/8 and GOV/2008/59 — two consecutive reports for Iran’s inventory as of November 2008. The IAEA report of 19 November 2008 states that from February 2007 to 7 November 2008 “…based on the operator’s daily accounting records, Iran had produced approximately 630 kg of low enriched UF6 [uranium hexafluoride]. All nuclear material at FEP [Fuel Enrichment Plant], as well as all installed cascades, remain under Agency containment and surveillance.”

The next IAEA report, released just last Thursday 19 February 2009, contained the results of the IAEA physical inventory verification (PIV) that took place between 24 and 26 November 2008. According to the PIV, Iran had produced 839 kg of low enriched UF6 from February 2007 to 17 November 2008.

The most important point is that these two “inventories” were very different animals. The 7 November inventory was based on adding up the entries in the operators’ logbooks and developing an estimate of overall LEU stockpile. For the PIV, the IAEA actually measured how much LEU the Iranians had on hand. Ideally, the two should match quite closely, so one would expect that the November LEU according to the two reports would be the same. If there is a discrepancy, then Iran may be hiding something.

Is there a discrepancy?

So if there is an extra 209 kg of UF6, why doesn’t the IAEA report scream fraud? Because according to the IAEA reports, all the quantities of UF6 produced balance out with the measured inputs so no uranium has gone missing and none that was hidden has suddenly appeared. All of the uranium and enrichment equipment was under constant IAEA monitoring so nothing was “discovered” except in a bookkeeping sense. The Iranians knew during the 7 November inspection that the IAEA measurement verification was coming a fortnight later. The Iranians are at least months away from getting significant quantities of highly enriched uranium (HEU), so it seems unlikely that they were trying to hide any LEU inventory for only a couple of weeks to get any head start on the IAEA inspectors.

The logbook inventory and the measurement inventory took place ten days apart. Could that account for more the difference? Most likely not. In the following measurement period, from 17 November 2008 to 31 January 2009 ( 75 days) the Iranians enriched an additional 171 kg of UF6 or 2.28 kg UF6/day. Similarly, from 30 August 2008 to 7 November 2008 Iran enriched an average of 2.17 kg UF6/day. So from November 7th to the 17th,we might expect them to enrich an additional 22 to 23 kg, or only about ten percent of the 209 kg difference we are concerned with. So the additional 10 days’ production cannot account for the difference.


It is more plausible that the Iranians, novices at the centrifuge business, have not developed sophisticated measurement and recording systems to track their own production. The IAEA reports do not contain enough detail to determine the source of the discrepancy but there are many possibilities. We do not know what containers the Iranians use to store their UF6 (the international standard is a steel cylinder called the 30B, which can hold up to 2.5 tons of UF6 and is too big for Iran’s program).  UF6 is a solid at room temperature and pressure. It is stored by pumping the gas from the centrifuges into a cylinder, where the UF6 condenses into a white solid. Because the gas condenses on the inside, the gas pressure does not go up as more gas is pumped in. Knowing how much UF6 is in the tank is not as simple, therefore, as reading a pressure gauge –the mass of the cylinders must be measured. The storage containers can be weighed continually; that is, the cylinders essentially sit on a scale but perhaps the Iranians don’t do that. If cylinders are moved from the production area to a storage area only when full and are then weighed, jumps in measured inventory are expected.

The 17 November PIV was the second IAEA physical inventory since Iran began enrichment in February 2007.  The first one was done in December 2007. The IAEA report (GOV/2008/4) announcing the results of the December 2007 PIV did not report any inconsistencies in inventory or Iran’s technical problems reporting LEU. The report was also the first time that enriched amounts of UF6 were disclosed and it commented that “the throughput of the facility had been well below its declared design capacity.” There are no prior published logbook enrichment amounts to compare the December PIV results to, probably because of the small cascade throughput. Consequently, we do not have evidence that Iran has ever understated production previously and obviously we have no comparison for historical logbook-to-inventory accuracy.

Not Enough for a Breakout…Yet.

By “breakout” they mean that, by starting with LEU, Iran could use its enrichment capacity to more quickly produce a bomb’s worth of the highly-enriched uranium needed for a bomb than it could if it were starting with natural uranium. While true in theory, the numbers presented by the IAEA indicate that Iran is not quite there.

Natural uranium has two isotopes (well, a trace of a third, but we can ignore that). In natural uranium, the U-238 is 99.3 percent and the U-235 is 0.7 percent. A reactor requires uranium that is 3-5 percent U-235 and a bomb requires uranium that is 90 or so percent U-235. By saying that the Iranians have enough LEU to further enrich to make a bomb, both articles are depending on simplistic calculations contained in the ISIS summary that look solely at the amount of U-235 in the LEU and see that it adds up to the 25 kg that is estimated to be required of a bomb. But just taking the quantity of LEU and multiplying by the U-235 concentration does not work because not all of the U-235 is recovered.

It is amazing how helpful even a little arithmetic can be. Taking the IAEA number, as of last November, the Iranians had fed 9956 kg of natural UF6 into their machines. Natural uranium is 0.71 percent U-235 so 9956 kg of UF6 contains 47.6 kg of U-235. During this time, the Iranians produced 839 kg of 3.5 percent LEU UF6. Multiply that out and we discover that the LEU contains 20.0 kg of U-235, which seems to imply there is not quite enough U-235 for a bomb, assumed to require 25 kg or so. (Note that if the calculations are redone using the masses, not of uranium, but of UF6, then the result is more than 29 kg, which would be enough for a bomb if it were pure uranium, not UF6. We suspect there is some confusion in the uranium-to-UF6 conversions.) But note that that also means that 27.8 kg of U-235 ended up in the waste or “tails,” which is to say that the Iranians are throwing away 58 percent of the U-235 that goes into the their machines. If the Iranians just recycled their LEU through the existing centrifuges, they would still throw away roughly the same fraction of U-235 so, while the U-235 is, indeed, present in the LEU, the Iranians cannot get it out in a pure form as easily as these articles suggest.

These numbers also indicate that the tails have a U-235 concentration of .45 percent, which is about double the normal commercial tails concentration. In normal centrifuge operation, natural uranium is fed into a set of centrifuges and the product is slightly enriched in U-235 and the waste is slightly depleted in U-235. The product of the first group of centrifuges is fed into another group for further concentration and so one until the desired concentration is reached. The waste stream is not just thrown away, however, and other centrifuges work to reconcentrate the waste to recover as much of the U-235 as practical. The high concentration of U-235 in the waste suggests that the Iranians have made a decision to get as much enriched product as quickly as possible with a given centrifuge capability. This is probably bad news for the rest of the world because it is consistent with getting a small amount of bomb material quickly, rather than a large amount of nuclear fuel eventually.

Another way to look at the problem is to consider that the Iranians have two constraints here: one is uranium and the other is enrichment capacity. The latest IAEA report states that the Iranians produced an additional 171 kg of 3.5 percent LEU UF6 between 17 November and 31 January. That is 116 kg of uranium equivalent over a period of 75 days. If the tails are still .45 percent U-235, that works out to 391 kg Separative Work Units (SWUs) or 5.2 kg-SWU/day.

If the Iranians had enough 3.5 percent LEU to feed into their machines to go for HEU, then, with 5.2 kg-SWU/day capacity, they could get 25 kg of 90 percent U-235 in 214 days if they left their tails at 1.2 percent U-235. However, they currently do not have enough uranium for this option. If they reduce the tails to 0.2 percent U-235, they will extract more U-235 and stretch their uranium supply, but then the enrichment to HEU of enough material for a bomb mass would take 384 days.

It’s Bad Enough, Don’t Make It Worse

We do not want to seem to be apologists for Iran. Their uranium enrichment program makes no economic sense. It could be consistent with a nuclear power fuel program but it is also consistent with a nuclear weapons program. It seems undeniable that Iran wants to at least maintain the option of developing a nuclear weapon. An Iranian nuclear weapon would be a danger to the world, and to Iran. We believe the rest of the world should work hard to avoid such a development but the world should develop policies based on the best analysis available. The hard facts are bad enough, there is no need for exaggeration.

8 thoughts on “Iran’s Uranium: Don’t Panic Yet.

  1. We received a thoughtful comment that in our breakout scenario discussion, we considered only the LEU produced up to 17 November 2008.

    From 17 November 2008 to 30 January 2009, Iran produced an additional 171 kg UF6 according to logbook entries (GOV/2009/8). We will assume that this amount maintains the average concentration measured by the PIV up to 17 November 2008 of 3.49%. (Iran has previously declared enrichment levels of up to 4.8%) The 171 kg UF6 contains 116 kg U and 4 kg of U-235. This gives a total of 23.8 kg of U-235 contained in the 1010 kg UF6 stockpiled at Natanz as of January 30th.

    Keeping in mind that much less of the 23.8 kg will be extracted if it is run through the cascade again, the amount is definitely not enough for a breakout scenario.

  2. “breakout capacity” is nonsense. Right now, Malawi has enough uranium atoms stored in its mines that it “could” one day decide to enrich at “hidden” facilities whose existence the IAEA cannot disprove. So, Malawi is a “virtual” nuclear weapons state…right?

    Give me a break. Iran is doing the same thing as Brazil and Argentina, with more transparency and international monitoring. The issue is not “nuclear weapons” but control over the nuclear fuel cycle which some countries are trying to monopolize. THis has been an ongoing battle between developing and developed nations for many decades.

    And, Melissa Fleming of the IAEA explained the “discrepancy”:

    “The (IAEA) has no reason at all to believe that the estimates of LEU produced in the (Natanz) facility were an intentional error by Iran. They are inherent in the early commissioning phases of such a facility when it is not known in advance how it will perform in practice,” said IAEA spokeswoman Melissa Fleming…”


  3. Lets get serious. Whether or not Iran is developing a nuclear weapon the country’s track record speaks for itself, as does its maniacal president.
    The country gives millions of dollars to terrorist organizations that want nothing more to kill anyone that does not share the same idealogy as them.
    This is all getting lost in the debate.
    My point is, knowing Irans track record, how can you trust them?
    WE MUST NOT HAVE A NATION THAT BEHAVES THIS WAY HAVE THE CAPACITY TO BUILD A NUCLEAR WEAPON. Let them first prove to the world they are responsible enough to have nuclear energy.
    We have two options. Take care of this situation now or continue to debate whether or not they are building a weapon. Now when they develop the weapon and Iran aims one at Israel or Southern Europe…what will we do? Might be too late at that point. This is a risk I am not willing to take.
    And understand this…Israel isnt either. The are carefully monitoring situation. If the USA and rest of world wont do anything, you can bet Israel will. Israel actualy requested bunker busting bombs from Bush right before he left office which was denied…what do you think these were for??

  4. To gary: Until you first mentioned “iran” I didn’t know on whose side you are.

    So I want to cite you:
    > My point is, knowing .. any state’s .. track record, how can you trust them?

    Yes, how could Iran trust the US that they suddenly do not claim that Iran has WMDs? I think there was a recent incident, where a country was attacked with this “excuse” – without ever finding any WMDs of course – hm what country was that again? I think it was quite near to Iran …
    According to your logic, the US really shouldnt have the ability to build nuclear weapons.

    After reading the article it should be obvios that diplomacy should still be the preferred choice. — not to mention that the US can’t afford another war like Iraq. And the US will be quite alone in this war. As it would be a war of aggression like in Iraq, the UN will again not help (which I totally understand and agree).

  5. I am very concerned about some of the technical details and assumptions contained in the essay.

    1) A minor point – Iran would have roughly 1075 kg of LEU (hex) by now.

    2) The essay assumes as valid that the needed U235 is the IAEA’s Significant Quantity – 25 kg contained U235. This number is for an implosion weapon. This decades old number is invalid.

    A good overview is found in Cochran and Paine’s NRDC report:

    “The Amount of Plutonium and Highly-Enriched Uranium Needed for Pure Fission Nuclear Weapons”

    From the intro:
    “The IAEA’s official “SQ” values also form the basis for public, media, and policy making assessments of the bomb-making potential of nations or terrorist groups seeking to acquire nuclear weapons.
    Unfortunately, as shown in this report, the IAEA persists in using SQ values
    that are outdated, technically erroneous, and even dangerous…”

    Your essay assumes a need for 27.8 kilograms of 90% HEU containing 25 kilograms of U235.

    The critical mass of 90% HEU in a good reflector/tamper is only 15 kilograms (containing 13.5 kg U235).

    You can see this in figure 42 on numbered page 95 (page 101 of the pdf) at the FAS website:

    Iraq’s pre-Gulf War 1 planned warhead design using calutrons apparently was
    for a 15 kg HEU core.

    Your 27.8 kg – assuming an excessive conversion and machining loss of 20% – yields 22.24 kg 90% HEU – containing 20 kg U235. This is way higher than needed.

    It is also unlikely that much more than 10% (recoverable) losses would occur. The higher loss number typically refers to plutonium extraction. Unlike HEU extraction, Pu production is a “wet” process. The intensely radioactive spent fuel containing only a few percent Pu is dissolved and a series of materials are added and removed. Significant quantities of fissile is lost in the various waste streams and plumbing.

    On the contrary, the quite small dry batches of LEU (hex) are essentally
    cooked to drive off the fluorine. Losses should be minimal. The metal is then directly cast into a core – a tabletop operation.

    3) What this implies:
    To produce 16.5 kg of 90% HEU metal (after 10% losses) yielding 15 kg HEU requires 678 SWU with 1.48% tails.

    At 5.2 SWU then 678/5.2 = 130 days NOT your 214 – 314 days.

    In addition, at the beginning of February Iran had 1476 centrifuges spinning in vacuum, ready to be fed hex. A few days ago Iran stated that all were now “operational”.

    This implies an output of not 5.2 SWU/day, but 7.2 SWU/day.

    Therefore: 678/7.2 = 94 days to weapon core NOT 314

    This all matter very much. I think FAS should modify the essay that is posted. FAS is an important trusted resource and it is imperative that it tell the story correctly. The FAS is quoted here:
    The Federation of American Scientists argues that experts using the 25 kilogram figure fail to take into account that — even if there is that much bomb-quality uranium mixed into their stockpile — not all of it is recoverable through enrichment methods.
    “The 12 to 13 kilograms they could produce would not be enough for a bomb,” says FAS vice president Ivan Oelrich vice president of the FAS Strategic Security Program.

    Moreover, the process of making the first uranium metal warhead from enriched uranium can lead to material loss, further reducing the amount left for a nuclear weapon.

    This is misleading and it is important that FAS correct its webpage AND contact that reporter.

  6. Yale Simkin in the above comment makes several excellent points. My impression is that there is fairly wide agreement that the IAEA numbers for “Quantities of Concern” are too high although what the numbers should be are not agreed upon. What is clear is that when certain thresholds are crossed depends both on an estimate of how much material a country has and how much material is needed to do something we would worry about. Or how much we should worry; there is a range of estimates about how many weapons the North Koreans might have because there is uncertainty about how much plutonium they have and also uncertainty about how much plutonium they need per weapon.

    Several things are going on here mentioned in the comment above that contribute to uncertainty. the NRDC report cited above considers implosion weapons. The critical mass of U-235 is, indeed, MUCH smaller for an implosion bomb than for a gun-assembled bomb. It may be that the Iranians are planning to use an implosion bomb. Or perhaps they may go first to a gun-assembled bomb and then, as they are more confident, they could move to an implosion device. It is often assumed that a country taking the uranium route is working toward an initial device that is gun assembled. (Note that if they take this second route, gun-assembled then implosion, then by the time they have an implosion design, they could also have plenty of uranium for a gun-assembled bomb anyway, so the distinction matters little.)

    If going to the trouble of making an implosion device, why not go the plutonium route? Plutonium has many other advantages, a smaller weapon primarily but also making plutonium in a small reactor is probably easier than a centrifuge program (although this is becoming less obvious as centrifuge technology advances). Of course, a small reactor, such as the North Korean model, is harder to pass off as part of a peaceful commercial program and that may figure prominently in Iran’s calculation. Also, with greater knowledge and computer power, designing an implosion device is constantly becoming easier and less risky. So a country may chose to go the uranium route AND the implosion route, skipping gun-assembled designs entirely. We agree that that is not a possibility that can be eliminated and Iran has reportedly done some experiment consistent with implosion devices.

    The second point, that with a good reflector, the uranium critical mass is much smaller than 25 kg even for a gun assembled device is also true. In practice, however, a gun assembled device would use more than the bare minimum critical mass. A nuclear explosive device works by creating a critical mass, either through compression or assembly, but as soon as that happens the fissionable material heats up and starts to expand, quickly destroying the criticality and the reaction stops. As a result, designers want the mass to be, not just barely critical, but as supercritical as possible when the reaction starts. That means that, ideally, a gun assembled bomb will not bring together two half critical masses to form a critical mass but two 99% critical masses so we start with two critical masses. (In fact, with the cylinder and slug design of the Little Boy design, the cylinder, because of its more open geometry, could already be greater than a spherical critical mass and the total mass was greater than two simple spherical critical masses.) Even so, a simple gun-assembled device is extremely inefficient, using only a couple of percent of the uranium. One might argue that a designer would hold out for more uranium rather than rush production of a bomb that was just barely critical. On the other hand, one might also argue that a bomb just barely critical might have a yield of “only” a kiloton and that is a worryingly big explosion.

    We were arguing in our post about the original press reports that misrepresented how much uranium Iran has and how quickly they could go from LEU to HEU and we stand by all those numbers. Simkim’s excellent comment raises an additional important question of what quantities of concern ought to be. This discussion of critical masses (perhaps in an abbreviated form) should have been included in the original post rather than simply citing the IAEA number. And we also agree that the IAEA numbers are too high but citing barely critical masses is too low. We also recognize that, with implosion devices, what is just barely critical is a strong function of the sophistication and skill of the bomb designers and builders.

    The last quote above about material being lost in processing is not from our report but from an AP reporter’s article. We talked to the reporter about losses during enrichment and did see an early draft of his piece that included the machining lose statement and we pointed that out as an error or, at best, a red herring, but it survived in the final article. We can’t take any responsibility for that.

  7. An excellent discussion on the misleading claims that Iran has achieved breakout capability has been done by Scott Kemp and Alexander Glaser from Princeton’s Program on Science and Global Security:

    They reiterate our point that as of January 2009, if all the U-235 were extracted from the 1010 kg UF6 stockpiled in an “ideal, loss-free cascade”, Iran would have “almost one IAEA ‘significant quantity’ “. Since this is not realistic, they discuss other breakout scenarios to conclude that it would take Iran a year and more realistically three years to produce a significant quantity of HEU.

    An interesting point in their conclusion – if uranium hexafluoride is converted into uranium dioxide (a crystalline powder), then it would be impossible to further enrich it. UF6 is converted into UO2 in the process of fabrication, a natural step in the production of fuel for nuclear reactors.

  8. Itseems to me that some of this discussion is confuned. Using their figures they find that the Iranians at the time of this being written have produced about 20 kg of U-235 in LEU hex. They then say that this is not enough for a bomb. No amount of uranium in LEU is enough for a bomb. You do not extract the U235 from the LEU hex you extract the uranium which contains the 3.5% uranium 235.

Leave a Reply

Your email address will not be published. Required fields are marked *