Maybe you remember some earlier discussion of Iran’s new-generation centrifuges. The IAEA introduced the subject to us back in Feb. 2008:

44. On 8 November 2007, Iran stated that it “agreed that exchanging of the new centrifuge generation information” would be discussed with the Agency in December 2007 (GOV/2007/58, para. 33). On 13 January 2008, the Director General and Deputy Director General for Safeguards visited an AEOI R&D laboratory at Kalaye Electric, where they were given information on R&D activities being carried out there. These included work on four different centrifuge designs: two subcritical rotor designs, a rotor with bellows and a more advanced centrifuge. Iran informed the Agency that the R&D laboratory was developing centrifuge components, measuring equipment and vacuum pumps with the aim of having entirely indigenous production capabilities in Iran.

So, just how indigenous are these designs, exactly? Do they have an identifiable lineage?

For glimmerings of insight into this question, we must turn to the indefatigable Mark Hibbs, who has had a few articles on related subjects in recent years.

According to MH in the Jan. 29, 2007 issue of NuclearFuel, Pakistan developed not just the by-now-familiar P-1, based on URENCO’s SNOR and CNOR machines, and not just the P-2, based on URENCO’s G-2, but also a P-3 and P-4, starting in the mid-1980s.

(As an aside, this might explain why Khan and his associates were prepared to start selling the earlier technology: it was obsolete. Old machines were being replaced, and those elements of the supply chain that were devoted to their components and materials (e.g., CNOR bearings) would no longer serve a purpose for Pakistan, KRL, or Khan — except to make some money.)

Hibbs writes:

While individual segments of Pakistan’s aluminum P-1 model had a throughput of less than 1 SWU/year, P-2, which features two maraging steel rotor tube segments, had a throughput of about 5 SWU/yr. P-3, the first of two later centrifuges, according to the intelligence information, is a four-tube model with a throughput of just under 12 SWU/yr. A successor model, P-4, may have a throughput slightly over 20 SWU/yr, the information indicates.

It should come as no surprise, by now, that the P-3 and P-4 were based on URENCO models, the four-tube 4-M and the six-tube TC-10, also known as SLM, according to MH in the Feb. 15, 2007 issue of Nucleonics Week.

It’ll certainly be interesting to learn how much the IR-4 — apart from having carbon-fiber rotors — resembles the P-3. Certainly, to judge by what Hibbs writes, the IAEA had its suspicions.

Another possibility, though, is that the IR-4 is basically an upgraded, carbon-fiber version of the P-2, owing nothing in particular to the P-3 or P-4.

As for the “more advanced centrifuge” mentioned by the IAEA last year, we’ll just have to wait and see.

Edited lightly for clarity.