Jeffrey LewisTech, Politics and Perspective On Iran

Paul has done a stellar job keeping up on the most recent IAEA report, making Total Wonkerr a go-to resource on Iran’s centrifuges.

The Washington Post’s Karen DeYoung has an excellent, balanced article on the technical and diplomatic implications of the report.

Anyway, now that we are a day out from the release of the IAEA report, I wanted to talk about the technical progress that Iran has made, what that means for diplomacy and the need for an unclassified NIE on Iran’s nuclear programs.

Is the Glass Half Full?

David Albright tells DeYoung that the ever-present metaphorical glass “is a little more than half full.” David and I did essentially the same calculation, so I thought I would be explicit about the assumptions regarding Iran’s centrifuge operations.

Based on the letter from IAEA DDG for Safeguards Heinonen and Paul’s awesome reporting, we can assume that Iran began feeding UF6 into the eight cascades on or about April 15, when most of us were filing our taxes.

Between April 15-May 22 (37 days, not counting the 22nd, or 888 hours), Iran fed approximately 260 kg of UF6 into the eight cascades. Assuming a feed rate of 70 grams an hour, eight cascades should consume approximately 500 kilograms of UF6. If Iran consumes just 260 kilograms, than the centrifuges are operating a little above 50 percent of what one would expect.

First, the Iranians may feeding hex only intermittently for a variety of reasons. Paul, for example, reported that the Iranians were being “cautious.”

Second, Iran might feed the hex more slowly to boost the enrichment levels. Note that the IAEA report claims Iran has enriched to 4.8 percent U-235.

Gary Samore told Chris Nelson that Iran’s centrifuges aren’t spinning at the full 350 m/s clip:

My sources tell me that the machines are NOT operating at full capacity (i.e. they are spinning at lower than optimal speeds to avoid crashing). Also note that the IAEA report on the amount of feel material processed thus far to produce low enrich uranium (260 kg of natural UF6) is a very small amount.

Iran might feed UF6into the centrifuges more slowly to compensate for the slower speed of the centrifuges, producing higher levels of enrichment but at a much slower pace.

This is a straightforward technical tradeoff—the Iranians have rushed ahead, meaning that the corresponding breakout capability is less than one might expect for the same number of centrifuges.
I should note that Nelson-san quote me, too, nodding approvingly at Gary’s bottom line:

Iran is making technical progress but still far from having a credible nuclear weapons break out option.

Missing the Cascades for the Centrifuges

Although I stick by that bottom line, a good friend of mine writes in to observe that the technical discussion “may be obscuring the most important thing.” Iran is “out of compliance with a forceful and unequivocal UNSC resolution. So we may be minutely examining technical points where the issue is really diplomatic.”

At any given moment, the sense of urgency that the international community ought to have, based on the present technical consequences of their non-compliance, is subject to debate. What is not subject to debate is that, left unchecked, they will eventually get to a point where the urgency is not uncertain at all.

So, How’s that Iran NIE Coming Along?

Our public debate would be improved immensely, by the way, if the Bush Administration would release the overdue, unclassified NIE on Iran’s strategic programs—as it is required to do so by law. Word is the NIE is almost done, but not quite.

On that subject, four Democratic Senator’s have written to President Bush, reminding him that the FY2007 Defense Authorization act required reports on our Iran policy (the one we don’t have, literally) and updated NIEs on relevant subjects by January 2007.

My favorite part is when the authors—respectively, the Senate Majority leader and chairmen of the Armed Services, Foreign Relations and Intelligence committees—asked if the President might at least estimate when the overdue reports might be submitted:

May 22, 2007

The President
The White House
Washington, DC 20500

Dear Mr. President:

We respectfully call your attention to an unfulfilled provision of law that would greatly enhance Congressional oversight of U.S. policy on Iran.

As you know, Section 1213 of the fiscal year 2007 Defense Authorization Act (P.L. 109-364) requires you to provide Congress with an unclassified and classified report on your policy objectives and strategy regarding Iran. Congress expects that the Section 1213 report will address questions that have been raised about the role in your strategy of U.S. diplomatic, financial, intelligence, military and other activities, and development, public diplomacy, democracy, education, cultural and other programs relating to Iran.

Section 1213 also requires the Director of National Intelligence to submit to the Congress updated and comprehensive national intelligence estimates (NIEs) on Iran, in both classified and unclassified form, and we understand that several Iran-related NIEs that should address this requirement are nearing completion.

There is great interest in the Senate in ensuring an effective US policy addressing the challenge of Iran. Since the January 2007 deadline in law for filing these reports has long passed, we would appreciate an estimate from you on the expected completion dates for these reports. Rigorous oversight by Congress is an essential part of crafting effective national security policy, and receiving and reviewing these reports as soon as possible is important to accomplishing this task. We look forward to your update, and in the upcoming months, we hope we can work together on crafting an effective policy response to this important foreign policy challenge.

Harry Reid
Carl Levin
Joseph R. Biden Jr.
John D. Rockefeller IV

P.L. 109-364



SEC. 1213.

(a) Submittal to Congress of Updated National Intelligence Estimate on Iran-

(1) SUBMITTAL REQUIRED- The Director of National Intelligence shall submit to Congress an updated, comprehensive National Intelligence Estimate on Iran. Such National Intelligence Estimate shall be submitted as soon as is practicable, but not later than the end of the 90-day period beginning on the date of the enactment of this Act.

(2) NOTICE REGARDING SUBMITTAL- If before the end of the 90-day period specified in paragraph (1) the Director determines that the National Intelligence Estimate required by that paragraph cannot be submitted by the end of that period as required by that paragraph, the Director shall (before the end of that period) submit to Congress a report setting forth—
(A) the reasons why the National Intelligence Estimate cannot be submitted by the end of such 90-day period; and

(B) an estimated date for the submittal of the National Intelligence Estimate.

(3) FORM- The National Intelligence Estimate under paragraph (1) shall be submitted in classified form. Consistent with the protection of intelligence sources and methods, an unclassified summary of the key judgments of the National Intelligence Estimate should be submitted.

(b) Presidential Report on Policy Objectives and United States Strategy Regarding Iran-

(1) REPORT REQUIRED- As soon as is practicable, but not later than 90 days after the date of the enactment of this Act, the President shall submit to Congress a report on—
(A) the objectives of United States policy on Iran; and

(B) the strategy for achieving those objectives.

(2) FORM- The report under paragraph (1) shall be submitted in unclassified form with a classified annex, as appropriate.

(3) ELEMENTS- The report submitted under paragraph (1) shall—
(A) address the role of diplomacy, incentives, sanctions, other punitive measures and incentives, and other programs and activities relating to Iran for which funds are provided by Congress; and

(B) summarize United States contingency planning regarding the range of possible United States military actions in support of United States policy objectives with respect to Iran.


  1. Allen Thomson (History)

    > the overdue, unclassified NIE on Iran’s strategic programs

    I will be very interested to see what it has to say about Iranian acquisition of longer-range ballistic missiles, ICBMs in particular.

    As far as I’ve been able to find, there is no, zip, zilch public information indicating that the US has evidence that Iran currently has ICBMs under development. The statements are that Iran could have ICBMs by 201x, but that appears to be based on estimates of general technical capability.

    And then there’s the BM-25…

  2. yale (History)

    Jeff, Your estimates appear to carry the assumption that Iran (amazingly foolishly) started pushing 70 grams of UF6 per hour into machines right from spin-testing.

    Don’t know about you, but if I were going from “operating” them as the April IAEA letter describes, to full “production”, I would take AT LEAST 2 weeks or more of slow build-up per cascade.

    Remember, the April letter described “some UF6 is being fed into those cascades”, without stating even which ones how many or how much at what rate – altho it implies quite tentative feeding.

    Based on good engineering practice, the Iranians appear to have achieved incredible performance and in record time.

    What is the basis for your assumption?

  3. John Field (History)

    Now, wait a minute…

    4.8% enrichment and 0.4% tails would be 0.41 SWU/kg input. 70 gms/hour for a year is then 250 SWU/cascade which is then 1.5 SWU/machine. And, the Iranians are only feeding half that amount, and the enrichment level was only stated to be as high as 4.8%, so we could be talking about them achieving as low as perhaps 0.5 SWU/machine, all things considered. But, their machines were reported to be capable of 2-3 SWU/yr as I recall – a factor of 4 more than achieved.

    Now, if I were the Iranians, it seems to me that I would be interested in mapping out the separation factor versus flow rate for my centrifuges and the cascade. So, if my cascade is designed for 3.4% or so maybe and I wanted to get 4.8% out of it on my little test drive, I would have to run my feed rate 2/3 as fast just on a single machine basis to keep up with the SWU increase. And then, there is the problem that the cascade feed to tails ratio is no longer right for that higher enrichment level. So, it seems pretty likely that I’d end up feeding at about 1/2 the design rate if I wanted to enrich to 4.8%.

    This takes care of a factor of two, but not yet the factor of four observed. But, maybe the centrifuges are spinning at 300 m/s instead of 350 m/s like Paul suggested. Then, SWU would be half and everything would be about right.

    Once you’ve got hex in the centrifuges, why do you care how much you are running through the system? – since the mass of UF6 is relatively small compared to the rotor.(Unless maybe the molecular pump or seals are leaking like mad?) And, once they are spinning, it seems like you surely don’t want to shut them off because then you have to drop down through the rotor resonances as they slow down. And, the needle ball lubrication will not be good during the shutdown/restart because the rotor isn’t spinning fast enough to pump up the pressure.

    To me, it just makes the most sense that the centrifuges are running a little bit slow for safety, and the Iranians are mapping out the performance curve at the lower speed.

  4. Haninah (History)

    Two comments on the WaPo article:

    1) ”…recently increased Iranian support for the Taliban in Afghanistan.” Wha? Did this surprise anyone else? Before we invaded Afghanistan, Iran and the Taliban were sworn enemies – so much so that some people point out that by toppling the Taliban, we were doing almost as much of a favor to the ayatollahs as we did by toppling Saddam. Have things really gotten so bad that now the Iranians are supporting the Taliban?

    2) “But the level of enrichment—less than 5 percent—is substantially lower than the 90 percent required to make a nuclear weapon.” Come on, Jeff, where were you on that one? We all know that’s factually untrue. Ninety percent is the threshold for weapons-grade uranium, and is optimal for making nukes – but anything above twenty percent is HEU, and can be used to make a nuke.

  5. Jeffrey Lewis (History)

    Yale, you can disagree with my assumptions, but “incredibly foolish” earns you a time-out.

    The assumption is based on Heinonen’s claim that during DIV on April 15th at the FEB, Iran notified the IAEA that “some UF6 is being fed into those cascades.”

    There are polite ways to engage in discourse. A colleague just wrote to say “After reading the just posted piece I realize that I had assumed a different measurement date for the amount of UF6 fed. I had assumed 13 May, the date of the only known inspector visit to the site, as the date of the measurement. You assume 22 May. Why, out of curiosity? It moves the actual/optimal UF6 feed to about 50 from about 70 percent.”

    That’s a model for trying to work through the assumptions without being a jerk.

    Haninah — Matt makes the same point about an error in Langewiesche’s book. That’s the sort of simplification that I usually let go. As you know, the amounts scale such that a device becomes extremely difficult to deliver — making it less interesting to me. I suppose I’ll be wiping egg off my face when some rouge state slaps together a massive bomb with 20 percent HEU and delivers it with a ship.

  6. yale (History)

    Jeff, Jeff, Jeff!You misread my postI WAS NOT SAYING YOUR ESTIMATE WAS “AMAZINGLY FOOLISH”! I was describing -> THE IRANIANS

  7. Jeffrey Lewis (History)

    Sorry, sorry.

    Damn, I am cranky lately.

    You can tell I am exhausted.

  8. Jeffrey Lewis (History)

    You know the saying “One is an accident, two a coincidence and three is a pattern”?

    Well, I noticed the pattern—well before this episode.

    I’ve scheduled a vacation starting next Friday and I have guest bloggers coming in.

    So, not only do I realize that I am beat, I am taking corrective action. Can’t come too early.

  9. yale (History)

    Dr. J…Sorry for the ambiguity in my phrasing.

    To clarify:

    What I awkwardly said was that the Iranians would be foolish to go from coasting in vacuum to a full feedof 70 grams/hour of massive UF6 molecules.

    Cascades shattered in teensy shards of aluminium would result.

    I was suggesting that they would engage in a good practice extended Ramp-Up/Shake-Down process over, say, a couple three week period, before runningfull-tilt.

    As John pointed out before, they would be looking to create performance and stability curves.

    That is why I asked for your assumptions, to see if it included the predictable low and intermittant output during cascade rollout after vacuum. The periodbetween in “operation” and in “production”

    How the Iranians were doing startup would affect the analysis.

    In any event,

    1) 2100 centrifuges available by

    2) July 2007 with your calculated duty cycle of

    2) 50% and at say

    3) 1.75 kq-swu/y/centrifuge

    4) yields 20kg 90% HEU

    5) August 2009———————————————————That being said, leave all this crap behind, grab a case of Everclear and Southern Comfort and de-Wonkify for a couple of weeks.

  10. eaton

    My curiosity was piqued by the IAEA’s parenthetical comment that Iran has “since disconnected” one of its 164-machine cascades at the PFEP. Are these cascades being dismantled and moved elsewhere; to the FEP or another undisclosed location? Since the PFEP was never completed to its full potential of some five cascades it is curious that the Iranian’s have apparently begun to dismantle their test-bed facility and it would be interesting to know the motivations behind this apparent action. One significant possibility would be that Iran has reached the limit of its centrifuge production capacity and is cannibalizing existing cascades in order to complete those in the FEP. This could also be a frugality measure, a security measure to move components to a less vulnerable site, or an indication that Iran feels that there is no further scientific or technical knowledge to be gained from a separate “experimental” facility. Does anyone know the back story to this Iranian move?

  11. Andy (History)


    It hasn’t been reported on extensively by many in the press, but it’s become increasingly apparent that Iran has expanded its proxy war to Afghanistan. Herat, historically one of the calmest provinces and an area with extensive Iranian influence, has been a hotbed of Taliban and insurgent activity this year. Here’s one article on the subject I ran across recently:

  12. Jeffrey Lewis (History)


    The new ISIS report notes that:

    Iran is not using the pilot fuel enrichment plant at Natanz to demonstrate such a capability. This plant remains unfinished, and the IAEA report states that one of the two cascades in the pilot plant has been disconnected. Contrary to expectations, Iran did not use the pilot plant to demonstrate the simultaneous operation of cascades. The reason for this is unknown but there are several possible explanations. One is that Iran is sufficiently confident in its abilities to operate the cascades that it felt it could safely forgo the pilot plant stage; another is that it is choosing to use the FEP as a pilot plant. In this way, Iran has the opportunity to orient itself with larger-scale cascade operations involving an entire module, or block of centrifuges. It is also possible that Iran is taking a technological risk, jumping as quickly as it can into industrial-scale operation without an adequate testing phase, and that the speed with which it has installed centrifuges at the FEP will later prove problematic.

  13. hass (History)

    Andy -Herat has been calm BECAUSE of Iranian influence which includes building roads there. An anonymous claim without evidence of Iran backing the Taliban (which Iran opposed for years when the US was flirting with them) doesn’t make anything “increasingly apparent”