Tuesday, April 10, 2007

Iran WarWatch (Part Eight)

I wrote the following as a posting in the Stratfor Iran Forum after the Iranians made their big claim about "industrial scale processing" for uranium enrichment.

Color me yawning. This kind of talk will be used by the West as part of the case to whack Ahmadinejad when the war that I expect finally comes, but as a real threat, it is way down on the list.


Iran Says Nuclear Enrichment Reaches Industrial Scale
by Marc Wolfensberger and Patrick Donahue

April 9 (Bloomberg) -- President Mahmoud Ahmadinejad said Iran has begun enriching uranium on an industrial scale, stepping up the country's defiance of the United Nations and risking an escalation of tensions over its nuclear program. ``I am proud to say that right from today our country has entered the group of countries that produce nuclear fuel industrially,'' Ahmadinejad said today at the Natanz uranium- enrichment site. While the president didn't specify the scale of enrichment, Iranian nuclear negotiator Ali Larijani confirmed to reporters that uranium gas was being fed into 3,000 centrifuges. The International Atomic Energy Agency said Feb. 22 that Iran planned to have 3,000 centrifuges at Natanz by May, though a UN official with direct knowledge of the IAEA's Iran inquiries at the time called that goal optimistic. About 1,500 centrifuges spinning non-stop for a year would be needed to produce the 28 kilograms (62 pounds) of 90 percent-enriched uranium needed for a bomb, said nuclear physicist David Albright of the Institute for Science and International Security in Washington.

The output of a centrifuge is based upon the rotational speed of the drum, along with the height and diameter (and a few other parameters, I suggest Cochran and Tsoulfanidis' The Nuclear Fuel Cycle: Analysis and Management, pg . 65-67 for a basic treatment).

It is estimated that the Iranian centrifuge design (based on a P1 type centrifuge) can churn out about 2 to 2.5 kg-SWU per year, centrifuge (where SWU is a Separative Work Unit - a term used in uranium enrichment). Numbers as high as 3 kg SWU/year have been touted for the type of centrifuge, but that is for one in optimum condition, run by experienced operators and with access to high-quality maintenance and parts.

So, if we are to believe President Ahmadinejad, then we have Iran cranking away with 3000 centrifuges. If he wants to go to 90% enriched for a bomb (instead of the civlian grade stuff he claims), then he needs around 30 kg of HEU. Note that that 30 kg figure is the low end - and does not take into account the HEU that might be lost in production, machining, etc.

How many kg-SWU will be required to get 30 kg of 90% enriched HEU? You can pencil it out, or you can use a Uranium SWU caculator provided by the FAS. I used a tails assay of 0.3% and got 5788 kg-SWU to get it. So... it does seem to pencil out to be around a year or so to get 30 kg.

Now, I will also grant you that the Iranians may have a reasonably modern weapons design with excellent neutron reflection capabilities and thus might be able to get by with just 30 kg at 90%, but I'll believe it when they test it successfully. Remember, Little Boy had 50 kg of 89% enriched material, along with a tamper of 14 kg of 50% enriched and the first weapon the Iranians will make will be similar to that (Trinity and Fat Man were both Pu weapons and Iran is not pursuing that route). Yes, I know they can boost it with various tritium producers to make it more powerful than Little Boy, but the uranium part will be very similar.

So we are looking at twelve to fourteen months at the earliest for Iran to have accumulated enough for a credible bomb design. They need all 3000 working, all the time. Now that boundary condition should be viewed as an extremely aggressive and unlikely when we consider other factors such as:

  1. What is the mean time between failures of these centrifuges? The Iranians, with all their skill, almost certainly do not have the industrial and metallurgical infrastructure to maintain their centrifuges in the manner you would see in Europe (the US will be building two centrifuge facilities, one at Piketon, OH and the other in Eunice, NM, but those will be years away). So even if we have two years or so MTBF the Iranians will have 5 or 10 fail per day.
  2. Can they maintain the ones that do fail? Bearings, rotors, etc. This is not trivial in a complex system such as a centrifuge cascade. The whole rationale for using centrifuges is based on its efficiency - which is reliant on speed of rotation. They can go easier on the system by easing rotation speed down to 300 m/s from 350 m/s, but that lowers the production rate. Good for the West.
  3. Steady supply of Hex. UF6 is a nasty substance. Think corrosion, think leaks, think potential for supply interruptions. No figures for you on this one, just another weak link in the chain.
  4. Then you have to do the metallurgy on the metal. Get it in the right phase, cast it, machine it, etc.
  5. Now, you have to test this beast you've built with all this time and effort - otherwise, no one will know for sure you are in the nuke club. So now you light it off, hope it explodes and...
  6. Iran now has to go through the whole process again. Of course, this all assumes the Israelis or the U.S. have not begun bombing by now.

My suggestion - don't stress too much for about two years, at least about an Iranian bomb. You get more time if they showcase stuff like pulling out low-enriched uranium to make fuel rods (they might do this if they decide to "prove" to the world they are in the business of the peaceful use of nuclear energy), giving them less material to continue the enrichment path.

Even two years is optimistic to the point of delusion for the Iranians. They have to "shake down" the centrifuges, apply "lessons learned" from the process, get their maintenance figured out, avoid sabotage by Western agents, provide enough UF6 to get maximum throughput, etc. I personally don't see them being able to do it in five years. Consider the two year mark a 'worst case' scenario.

No comments: