The Washington Post reported today on a new assessment from the Institute for Science and International Security (ISIS) on Iran’s progress – and problems – with uranium enrichment. (H/t: Ed Morrissey at Hot Air.) Readers are unlikely to mistake either piece for an assessment that Iran’s nuclear program isn’t a threat. But most readers are likely to miss the potential implications of the ISIS assessment when viewed in the context of other developments in the nuclear program. Iran’s apparent failure to address centrifuge problems at the uranium enrichment plant in Natanz could merely be evidence of technological incompetence or fecklessness. But it could also be a collateral indicator that Iran is placing a higher priority on activities elsewhere.
It could, in fact, indicate that Iran has progressed enough with undeclared processing sites that the declared operations at Natanz, and perhaps at the Esfahan uranium conversion complex as well, are not now the highest priority for resource allocation and troubleshooting.
The ISIS report as written is careful and sober, although a close reading reveals that its central attribution is circumstantial. There has, according to both IAEA inspectors and Iranian self-reporting, been a significant drop-off in the number of centrifuges in operation at Natanz. This decline was noticeable at different reporting dates in 2009, with the drop-off being assessed from a high-water mark in late 2008. The circumstantial aspect of the analysis lies in the apparent use of general anecdote from unnamed officials and industry experts about the difficulty of what Iran is trying to do, the likelihood of running into problems in execution, and the rapid pace at which the Iranians have developed their program. There are a couple of direct assertions that the Iranians are, in fact, having performance problems with their enrichment equipment, but the assertions are not specific or backed up with any detail.
I stress that this doesn’t mean I think the analysis is intentionally skewed. I don’t. But I do think it may have been performed with blinders on. And I am convinced that WaPo’s piece on it today conveys the ISIS assessment’s thesis so superficially as to be misleading. “Technical setbacks cause Iran to falter in push to enrich uranium” sends a very clear message; but the ISIS report actually considers possibilities other than technical setbacks to explain the decline in centrifuge operations.
It basically dismisses them, however, outlining a key missing piece of the puzzle as follows:
Finally, there remains the question of what happened in late 2008 or early 2009 that resulted in a significant lowering of LEU [low-enriched uranium] output. Any of the above problems could have risen to a level that drastically cut LEU production. In addition, an unknown event could have caused the drop.
So: Data Point 1. There was a significant decline in LEU output in late 2008 to early 2009.
A starting point for broader analysis is this element of ISIS’ own commentary on the last IAEA report on Iran, from November 2009 (emphasis added):
During the most recent reporting period, from August 1 to October 30, 2009, Iran produced a total of 255 kg of low enriched uranium (LEU) hexafluoride, bringing its total to 1,763 kilograms of LEU hexafluoride. The average daily rate of LEU production has increased to 2.8 kg per day, up slightly from its rate of 2.75 over the two previous periods. Because this LEU is being produced in 656 fewer centrifuges (a reduction of 15 percent), Iran’s P1 centrifuges appear to be operating more efficiently than in previous months, although little information is available about this subject.
Comparing the last two quoted passages with Figure 2 from the ISIS assessment, we discern that there was a precipitous drop-off of enrichment between August and December 2008, followed by a significant increase between December 2008 and April 2009, and the maintenance of a slight increase after that through October. Although the level of enrichment at Natanz has not been restored to its high from August 2008, the ISIS evaluation in November of last year was that the level of late 2009 was being maintained with fewer centrifuges, operating more efficiently.
Data Point 2. LEU output ramped back up, although not to the level of August 2008, and in 2009 was being produced more efficiently by fewer centrifuges.
It’s worth noting, without placing unjustified confidence ourselves in superficial observations, that this recitation of events sounds a little different from the “technical setbacks” narrative. It sounds like a change in operating profile between late 2008 and mid-2009, one that could as well have been deliberate as driven wholly (or even mainly) by technical problems. It’s certainly possible that Iran shut down a batch of poorly-performing centrifuges and has brought some better ones online, and that that’s all there is to it. But then again, whether we see that as a setback, or as progress, is really a question of perspective: glass half-empty or half-full.
The question this scenario begs is why the Iranians haven’t restored more centrifuges to operation (or why, as ISIS points out, there seems to be a discrepancy between the number of operational centrifuges given by Iran’s nuclear chief in December 2009 – 6,000 – and the number reported by IAEA in November, which was less than 4,000). One of the problems for us in making assessments is that we simply don’t know enough to confidently eliminate possible reasons.
That in turn means two things, however. It means we can’t absolutely eliminate the possibility that the Iranians are just bumbling along having difficulties, racing into things too fast and looking poorly prepared and inefficient in the process. But it also means we can’t eliminate the possibility that they are shifting their priorities elsewhere, and that the reason – whatever it is – that they don’t think it’s as important as outside observers would suppose, to maximize their LEU output at Natanz, makes perfect sense to them.
One clue lies in the drop in the feed of uranium hexafluoride (UF6) to Natanz, from the uranium conversion facility at Esfahan, in late 2008. Figure 3 from the ISIS assessment shows a substantial drop from August to December that year. The feed then increased again through August 2009, and declined again by October.
The sense at ISIS, that something significant to Iran’s level of uranium processing happened in the period late 2008 to early 2009, is bolstered by the drop in UF6 feed to Natanz in that same period. But what was it? Why did the converted uranium feed decline at that point?
Data Point 3. The UF6 feed from Esfahan to Natanz declined at the same time the production of LEU dropped significantly at Natanz.
The most important clue to that is probably the evidence, confirmed by multiple sources, that beginning in the latter half of 2008, Iran shifted her uranium feed entirely to uranium mined inside Iran. Until that shift occurred, Iran was processing yellowcake obtained from South Africa in the 1970s. She has developed two mining sites inside her territory in the past decade, however, one in central Iran at Saghand and one near the southern port of Bandar Abbas at Gchine (also rendered Gachin). Iran reported achieving the capacity to produce her own yellowcake in 2006. In November 2009, a Bloomberg report indicated that commercial satellite imagery showed a substantial increase in activity at the Gchine site, where facilities for separating and milling uranium are co-located with the mine itself.
Subsequent analysis at the expert blog “Arms Control Wonk,” here and here, endorsed the soundness of the Bloomberg report’s conclusion. The 10 November post contained a reference as well to a December 2008 article by nuclear industry expert Mark Hibbs in the journal Nuclear Fuel (dated 15 December 2008), entitled “All of Iran’s UF6 centrifuge feed now indigenously mined, milled.” (This article is not available through Google search but can reportedly be accessed by journal subscribers and via Lexis-Nexis.)
Data Point 4. Iran’s uranium feed for the conversion and enrichment processes shifted to indigenously-produced yellowcake at the same time LEU production dropped at Natanz.
The shift to indigenously-produced uranium was to be expected, given the limited amount of yellowcake Iran had on-hand from her Shah-era purchase from South Africa (less than 600 tons). Iran’s use of her own uranium hardly qualifies as a suspicious development in and of itself. But two factors justifying suspicion are nevertheless present: the history of the Gchine site, and its procedural exclusion from IAEA monitoring.
The practical import of the latter can be expressed most aptly as follows: we have no accountability regarding how much uranium has been mined and milled from Gchine, and then forwarded for conversion to UF6. As the Wonk posts suggest, we can make educated guesses about how much has been mined and milled, based on the expanding size of the reservoir for ore wastes. But we don’t have any reliable means of figuring out what has happened to it after that. Iran has reported a capacity to produce yellowcake, and has continued to feed Esfahan and Natanz. If Mark Hibbs’ piece is accurate, we can presume that all the feed material going into the enrichment process since late 2008 was indigenously produced. But how that amount compares to what has been mined and milled, we have no method of accounting for.
Iran has usable uranium now, in other words, outside IAEA accountability. The utility of the IAEA process has been predicated on accountability relative to the baseline established with the yellowcake from South Africa, and that accountability is no longer reliable.
Data Point 5. Indigenously-produced uranium yellowcake enables Iran to bypass IAEA accountability on her South African-supplied baseline.
The possibility of this day arriving has always been there, of course. The UN has known for years of Iran’s uranium mining program. But the mines and separation facilities are not subject to IAEA inspection, and analytical attention has been focused on the Saghand mining site since the early 2000s anyway, rather than on Gchine. As the links above indicate, Iran has been at some pains to encourage that focus, promoting Saghand with tours and briefs at international conferences, and making open declarations about the status of the associated milling site at Ardakan (Erdekan). In its July 2009 report, IAEA indicated observing activity at both mining sites (Saghand and Gchine) in satellite imagery – a rare check in the block for keeping track of these facilities. But, again, it does not inspect them.
As mentioned above, however, Gchine has a suspicious history. It was named in foreign intelligence documents outlining the “Green Salt” project, one of the efforts connected with Iran’s nuclear weaponization program, and was thought to have an association with the military in the 1990s. It was operated, moreover, by the shadowy private firm Kimia Madaan, apparently set up to evade sanctions in Iran’s nuclear procurement process, through the summer of 2003. In June of that year Iran’s Atomic Energy Organization (AEOI) assumed control of it; timing that was, to say the least, interesting, given that it was in that period that US intelligence assessed Iran to have suspended the nuclear weaponization effort (see the 2007 NIE).
The IAEA framework has never been foolproof, of course, but the public has been lulled into a sense that it represents a reasonably comprehensive form of accounting on Iran’s nuclear programs. Analysis from the last few months should convince us that it isn’t. Most people would never have heard about the dramatic increase in uranium mining at Gchine, which has been underway since mid-late 2008, without the Bloomberg report from November. Although IAEA was aware of it at some level, based on its July 2009 official report, the Agency has no responsibility for inspection or accountability at the mining sites – and therefore has had no chartered incentive to make those sites a centerpiece of any broader-scale forensic effort.
The key question remains whether Iran has the ability to perform the processing done at Esfahan and Natanz outside of those facilities, and to perform them on something approaching an industrial scale. Yellowcake has to be converted and enriched somewhere. The conversion of yellowcake into UF6 suitable for enrichment is done at Esfahan. Natanz is the site of the centrifuge cascades that enrich the UF6, currently to a low, sub-weapons-grade level. We don’t know of other sites in Iran for performing these tasks.
Massive tunnel complexes have been excavated adjacent to both sites in the last decade, however, and IAEA access to them has been almost entirely prohibited. Inspectors visited the tunnels at Esfahan in 2004, when excavation had been underway for a few months, and it should be no surprise that they were empty at that time. There has been no further access to Esfahan, and no visit to the tunnel complex at Natanz, where excavation was detected in 2007. Publicly-available estimates from 3-6 years ago about the size and capacity of these underground facilities must of necessity be outdated by now: to insist that they could not, in 2010, house any more than they were assessed to be capable of in 2004 or 2007 would be inherently faulty analysis.
It is also faulty to imagine, given the excavation revealed at Qom’s Fordo site in September, that the prospect of other undeclared facilities is unlikely. In fact, there is a list of facilities suspected of connection to Iran’s nuclear program that has gone almost entirely uninspected, including the Lavisan site outside Tehran, reported by the usually-reliable expat group “National Council of Resistance of Iran” (NCRI) to have been reconstructed near its original location after a complete razing in that busy summer of 2003. Lavisan’s activities have been thought to be related to weaponization research, along with those of Parchin, also located outside Tehran – neither is currently suspected of a role in uranium processing; but as a somewhat atypical New York Times story noted last month, continuing suspicions about sites in Iran are based on a record of secretiveness, bad faith, and surprises sprung by the regime in Tehran. The Iranian pattern of covert activities and the use of tunnels is a sound indicator that where there are tunnels, there is covert activity.
The break in Iran’s LEU production volume, in late 2008 and early 2009, might have been a very simple function of delays incident to bringing the indigenously-produced yellowcake online. And ISIS may be focused on the right issue: maybe it was all a matter of centrifuges breaking down and things going wrong. But the ups and downs of LEU production are also consistent with the beginning of a reallocation of feed material to a covert processing network. The timing matched the onset of Iran’s production of unaccountable source material too well for that factor to be dismissed out of hand. The subsequent establishment of a fairly regularized LEU output after the dip, especially if the output is maintained on a more efficient basis than previously seen, argues a reasonable level of control over the current processing regime. The latter half of 2009 doesn’t look so much like the Iranians being at their wits’ end with the darn centrifuges, as it does like the production of a deliberate amount of output.
As to the narrow question of why they are letting centrifuges sit idle at Natanz, possible reasons are certainly that their supply of UF6 feed material is iffy, or that the machinery has broken down. Analysts are pretty convinced that Iran’s supply of enrichable uranium is dwindling, and that she will be facing a shortage soon, a condition that is held to explain Tehran’s overtures to Kazakhstan, Guyana, Tanzania, and Venezuela for uranium cooperation and mining deals. Iran repeatedly claims to have at least 30,000 tons of uranium to mine from her own soil; but while 1970s-era surveys and OECD estimates put her uranium potential at around 25,000 tons, the great majority of that is thought to be difficult and even uneconomic to recover. So Iran may well be facing an effective shortage in the near term, even if she has plenty of uranium in theory.
But there is also the valid possibility that Iran’s priority is not, today, getting the production level back up to its high of August 2008 at Natanz. Bringing the idle centrifuges there online would cost money – and processing more UF6 there means submitting more to the IAEA’s ledger. The truth is, enriching less uranium at the declared facility at Natanz doesn’t necessarily mean less total uranium is being enriched.
The larger and more important point about all of this is that we simply don’t know. Now that Iran is mining and refining her own uranium for use in the enrichment process, we even more don’t know.