Joshua PollackIran: Compliance in Defiance?

Warning: Even more speculative than usual.

The Iranian government has spoken: in response to the the rebuke contained in Friday’s IAEA Board of Governors resolution, Iran has defiantly announced that it will build another ten enrichment facilities:

“We had no plan to build many nuclear sites like [enrichment facility in the central city of] Natanz but it seems that the West do not want to comprehend Iran’s message of peace,” Head of the Atomic Energy Organization of Iran Ali-Akbar Salehi said.

“The West adopted an attitude toward Iran which made the Iranian government to pass the ratification on construction of ten sites similar to the Natanz enrichment facility,” he added.

Very broadly speaking, this is the sort of escalation that many of us (including yours truly) had expected. But matters may not be as simple as they seem. There is another set of possibilities.

A Good News Story in Clever Disguise?

Recall the immediate context: a dispute between Iran and the IAEA about Iran’s obligation to report its decisions to establish new nuclear facilities. The Director-General’s last report to the BoG expressed suspicion about whether there were additional facilities yet to be declared:

14. The Agency further indicated that it still had questions about the purpose for which the facility [at Qom] had been intended and how it fit into Iran’s nuclear programme. The Agency also indicated that Iran’s declaration of the new facility reduces the level of confidence in the absence of other nuclear facilities under construction and gives rise to questions about whether there were any other nuclear facilities in Iran which had not been declared to the Agency.

This mistrust ran so deep that even after the Iranian side denied having additional facilities, the IAEA followed up with a letter asking, in effect, “Are you sure? Is that your final answer?”

16. Iran stated that it did not have any other nuclear facilities that were currently under construction or in operation that had not yet been declared to the Agency. Iran also stated that any such future facilities would “be reported to the Agency according to Iran’s obligations to the Agency”. In a letter dated 6 November 2009, the Agency asked Iran to confirm that it had not taken a decision to construct, or to authorize construction of, any other nuclear facility which had not been declared to the Agency.

That’s where matters stood on November 16, the date of the report. Friday’s BoG resolution appealed to the Iranians to respond to the letter:

5. [The Board of Governors] Calls on Iran to confirm, as requested by the Agency, that Iran has not taken a decision to construct, or authorize construction of, any other nuclear facility which has as yet not been declared to the Agency;

Well, now we have an answer to that appeal, after a fashion:

Upon the Iranian government’s decree, the AEO should begin the construction of five of the enrichment facilities over the next two months.

It should also propose locations for the remaining five enrichment plants within a two-month period.

“We have determined the location of five sites upon the president’s decree. Our enrichment sites will be constructed in the heart of mountains,” Salehi said.

Press statements are not the same thing as formal notifications, but the notifications should be arriving in Vienna shortly, if they haven’t already. If five sites really have already been selected, then undeclared plans to build them have been in place for awhile. Judging by how the Iranian side has framed its activities at Qom, construction may already have been underway for some time.

One way to see it, then, is that the Iranian side has seized the opportunity to get tough by coming clean, or to come clean by getting tough. In the two-level game of international diplomacy and Iranian domestic politics, this sort of Janus-faced response may be as close to a win-win outcome as ever happens.

How Many Enrichment Sites?

Not entirely by coincidence, the question of how many Qom-like enrichment sites Iran may have planned has been the subject of some wonk-jousting lately. Writing at the Bulletin, Ivan Oelrich and Ivanka Barzashka of FAS find Qom — which is due to hold 3,000 IR-1 centrifuges — too small to serve as a standalone HEU factory. They judge Qom (a.k.a Fordow) to be capable of producing roughly enough HEU for a bomb once every four years once it is at full capacity. This underscores the possibility “that Iran has a number of similar secret facilities—along with the required uranium hexafluoride conversion plants—and that Fordow is simply the only one to have been discovered… It also is possible that Fordow was the first of several planned secret facilities, and that, with its discovery, Iran has put further plans on hold.”

(This perspective appears close to the views of Gary Milhollin and Valerie Lincy of the Wisconsin Project.)

David Albright and Paul Brannan of ISIS dispute this view, judging Qom at full capacity to be capable of producing enough HEU for one to two bombs every year, depending on how capable the IR-1 machines really are. (Earlier this year, Ali Akbar Salehi put the figure at 2.1 kg SWU.)

Oelrich and Barzashka do not specify their assumptions, but may be using a SWU figure that is too low. Going by their previous analysis, they also may be using a 25 kg HEU figure per bomb, which is almost certainly too high. So on the narrow question of Qom’s potential, put me in the ISIS camp.

Whether Iran has been planning or preparing more hidden enrichment plants and other fuel cycle facilities is a separate question, though. We’ll probably have a better understanding in the coming weeks and months.

Why Was the IAEA So Insistent?

I don’t know, but my guess is, the IAEA Secretariat and the P5+1 knew something, and were hinting to the Iranians that they ought to come clean voluntarily, before the end of the year. After all, according to a story that appeared in the New York Times earlier this fall, the November 2007 NIE “listed more than a dozen suspect locations” in Iran. So the Intelligence Community seems to have had some pretty good leads for awhile now.

Recall that some of the more intriguing passages in the Key Judgments of the NIE relate to covert sites. The NIE defined “nuclear weapons program” — the thing it judged to have stopped in Iran in late 2003 — as “nuclear weapon design and weaponization work and covert uranium conversion-related and uranium enrichment-related work.” (Emphasis added.)

“Work,” in this context, presumably includes the planning and construction of facilities. Was that work resumed by the time that the NIE was issued? The answer is hedged: “We assess with moderate confidence Tehran had not restarted its nuclear weapons program as of mid-2007.” (Emphasis added.)

Finally, the Key Judgments include a view, also hedged, that Iran would return to building covert uranium conversion and enrichment facilities if it were to decide to produce HEU for weapons:

We assess with moderate confidence that Iran probably would use covert facilities—rather than its declared nuclear sites—for the production of highly enriched uranium for a weapon. A growing amount of intelligence indicates Iran was engaged in covert uranium conversion and uranium enrichment activity, but we judge that these efforts probably were halted in response to the fall 2003 halt, and that these efforts probably had not been restarted through at least mid-2007. [Emphasis added.]

If Qom is any example, the doubts expressed in the NIE had less to do with not knowing the whereabouts of the covert facilities, and more to do with doubts about the exact purposes of the facilities under observation. On that point, we might infer from the recent IAEA report and BoG resolution that there has been less doubt lately.

Viewed in this light, Iran’s “defiant” disclosure might be a voluntary foreclosure of its ability — probably already compromised — to use a network of covert sites to build the Bomb. If that’s so, then intelligence has secured some of the margin of security that negotiations seemingly could not.


  1. R. Scott Kemp (History)

    On the matter of P1 performance.

    Using validated hydrodynamic codes from the U.S. centrifuge program, Houston Wood has calculated that the maximum performance of a P-1 is 2.1-2.2 SWU/year. That would be consistent with Salehi’s remarks.

    However, throughput data published in IAEA BOG reports show that Iran’s centrifuge cascades have not operated at this level. The effective performance is closer to 0.6-0.9 SWU/year.

    In private communication with me, Albright and others have speculated that the low effective performance is because the cascades are not being operated continuously, implying that higher performance could be obtained if Iran really wanted it. This claim is not without difficulty since it begs the question why would Iran do this? They don’t suffer from a UF6 shortage, and it would not make sense to operate their cascades at a sub-optimal level if the goal is to accumulate as much LEU as possible.

    On the other hand, if the putative stopping and starting is due to technical difficulties, then the performance may well be about 2.1 SWU/yr when operating smoothly, but the number relevant to breakout calculations will still be their effective rate of 0.6-0.9. at least until those difficulties are resolved.

    I favor a different explanation: that the cascades are being fed near-continuously, but that they have high process losses to the vacuum system and from interstage mixing. This would be consistent with data from the 17 Nov 2008 PIV, which showed that 16% of feedstock wound up in the vacuum-system cold-traps and dumps. Interstage mixing could also account for some performance loss.

    Thus, I would predict that unless Iran re-designs the P1 and/or the 164-machine cascade layout, the effective performance will remain in the 0.6-0.9 SWU/year range.

  2. Yale simkin (History)

    Oelrich and Barzashka of FAS are using 27.8 kilograms of 90% HEU as the bomb quantity (way too much),
    0.5 kg-swu/year/centifuge (Nantanz-actual way too low),
    and (as I derive from their stated parameters) a tails of <0.2% (way less than Nantanz-actual).
    For the LEU to HEU case they use 0.5 kg-swu/year/centrifuge, 3.5% feedstock, and 1% tails (stated).

    Their numbers for both centrifuge performance (low) and HEU weapon requirements (high) strongly skew their results towards a long time-to-bomb.

    ISIS uses actual Nantanz performance numbers and even using their quite conservative tails numbers, get a seriously faster time-to-bomb.

  3. Josh (History)


    Thanks for this very insightful analysis.

    One issue that we may not be taking in account, when discussing the purposes of the Qom facility, is the possibility that it was not originally expected to hold IR-1 machines. Salehi also spoke of a “new generation” of 5 kg SWU machines, and was slow to announce what type were slated for Qom. Certainly, the official Sept. 21 account of Qom/Fordow as a “pilot plant” hinted at something other than the same old IR-1s. But the IR-2s, -2ms, -3s, and -4s may simply not be available yet. Just a thought, anyhow.


    Forgive my obtuseness, but can you tell me how you reconstructed those numbers from the text of the article?

    Oh, never mind. I see that the 0.5 kg SWU figure appears here, also in connection with the “four years” estimate. Thanks for pointing that out.

  4. Yale Simkin (History)

    The article you linked with the 0.5 kg-swu/yr/centrifuge only has some of Oelrich and Barzashka’s numbers. I reference this piece for their numbers (except where I specifically said they were derived).

    I find in Oelrich and Barzashka a pattern of strongly leaning towards minimizing the risk.
    I got into a bit of discussion with them, first in email which they reference in comment 1, and then much more irritably in comment 5 of an earlier posting at FAS here

  5. Josh (History)

    Thanks again, Yale.

    Another possibility we might consider, somewhat less rosy than the scenario above, is that the international community won a game of 3,000-Centrifuge Monty.

    In other words, Iran “prepared” a number of empty sites, then “allocated” one of them as a covert enrichment site.

    Under this interpretation, the unused sites were empty “discards,” or “holes in the ground,” as former IAEA DG Mohamed ElBaradei might put it. If this is so, we might not get to see the covert uranium conversion facility that could be expected to accompany a site like Qom, or the covert reprocessing facility that could be expected to accompany the Arak reactor.

  6. TNT
  7. PC (History)

    “Thus, I would predict that unless Iran re-designs the P1 and/or the 164-machine cascade layout, the effective performance will remain in the 0.6-0.9 SWU/year range.”

    Scott, I understand the cascades planned at Fordow will have a larger number of centrifuges than the 164-machine versions used at Natanz. Apparently Iran claimed they were experimenting with the larger cascades because of the layout of the facility, which was not originally designed for an enrichment plant (that rationale doesn’t make any sense to me).

    Could a larger number of centrifuges, however, address some of the performance issues you laid out?(unsure of the exact numbers but 180-190 seems a decent ballpark since the roughly 3000 machines will be divided into 16 cascades)

    Or perhaps it may be related to their intention to possibly use more advanced centrifuges?


    While I do suspect we will hear more about these multiple “contingency centers” Iran established, with others having possible relevance to the nuclear program, what seems to discredit the notion that this was some type of voluntary disclosure is the scale they’re talking about. Several more Fordow-sized plants, I can believe that. But several more Natanz plants? Show me the money. Of course, that’s not to say they could be planning additional Fordow sized plants anyway, but with this announcement, they would either need to back up their statements to the IAEA, or continue to say they don’t need to. Either option seems to work against them one way or another.

  8. J House (History)

    “…A growing amount of intelligence indicates Iran was engaged in covert uranium conversion and uranium enrichment activity, but we judge that these efforts probably were halted in response to the fall 2003 halt, and that these efforts probably had not been restarted through at least mid-2007. [Emphasis added.]”

    Despite all of the hedging in the NIE, the USIC was completely wrong on this count.

  9. BS (History)

    This may be the reason for the rejection of the recent proposals. Apart from a certain humiliation, any further evidence later presented of weapons enrichment activity would have had to be dealt within a purely non proliferation framework, Iran would have lost its excuse of fuel enrichment. Not only that – to have later withdrawn from the agreement would not have looked good for Iran – non transparency to new evidence and allegations , posession of more highly foreign enriched material and reimplementation of its enrichment activity. This train of thought leads to the unhappy conclusion that we will now witness an escalation on the issue, the premise being that Iran wishes to maintain full control of any covert activity. The possibility that Iranian material would have been witheld outside of Iran while answers to further new evidence were awaited is unlikely. It is possible however that a military strike on various of Iran’s facilities could have been timed to coincide with the fuel being outside of the country, making its return less likely.

  10. Josh (History)


    Thanks for calling attention to the FAS paper. The full-length paper is here. Haven’t had a chance to digest it yet, but will do so shortly.


    I have a hard time believing the Natanz-scale part, too. Despite what some Iranian officials have claimed, it’s not all-indigenous technology. (Zippe is spinning in his grave.) That’s a lot of parts and pieces that must be acquired somewhere.

    Pres. Ahmadinejad’s latest remarks seem to set up a conflict with the IAEA over disclosure and inspections. But this is just the opening bid.

    The Iranian claim that they had not started building any new nuclear facilities may be consistent with their apparent unilateral interpretation, per the Qom experience, of what it means to build a centrifuge facility, vs. a general-purpose bunker that may be “allocated” to the nuclear fuel cycle at some future time. If indeed they never intended to “allocate” and build more than one, the purpose of having many perhaps would be to intensify the “Kumchangri” problem for foreign intelligence agencies — an anti-simulation strategy, in effect. It’s an interesting dodge, from both a legalistic and a D&D perspective. But I’m just making an educated guess here.

    (Certainly, there does seem to have been some evasion of the specific question concerning the decision to build or authorization to construct nuclear facilities.)

    Following this hypothesis, the premature exposure of Qom would mean that the entire family of “holes in the ground” would now be eligible to provide domestic political benefits, much as Natanz did after its discovery. That could explain why they are being played up now, better than a desire to be transparent to outsiders.

    J House:

    By definition, if they hedged, they weren’t completely wrong. Can’t win ‘em all. But I’d say the IC was looking in the right direction.

  11. R. Scott Kemp

    Ref: PC Dec 2, 04:35 PM

    A different cascade layout could certainly help with some of the performance issues. The fact that they are larger doesn’t make much difference; what matters is the shape of the cascade and how well the match the isotopic abundance in the up and down flows from each stage. I don’t think you should regard a different design as meaning the performance would suddenly shoot-up to max. They are still losing a lot of material to the vacuum system and dumps, and that is independent of the shape of the cascade.

  12. Ivanka Barzashka (History)

    Thank you all for your interest in our technical assessment of Fordow. We have posted an updated Issue Brief (with prettier formating and clearer language), which is a technical appendix to our Bulletin article . We have carefully explained the origin of our numbers in a manner, which we hope is easy to follow even by non-technical parties.