Merlin: Both Russian Engineers Had Problems with the Blueprints
As I suggested in my response to Walter Pincus’ laundering of false claims about the evidence presented at the Sterling trial, DOJ went to great lengths to help CIA try to rebut claims made in James Risen’s book about the blueprints handed over to the Iranians.
Most importantly, by confusing the issue of “incomplete” plans with “flawed” plans, they suggested no one — not Merlin and not Jeffrey Sterling and not anyone else — should have a concern about the nuclear blueprints Merlin wrapped in a newspaper and dropped in a mailbox. They mean something very limited by that: that no one identified the flaws CIA purposely put into the plans to ensure they wouldn’t work.
There are several concerns with that:
- Merlin’s concerns, both nuclear and operational
- Human Asset 2’s concerns
- Problems with their logic
I’m going to bracket Merlin’s concerns for the moment (DOJ seems unenthused about answering my questions on where I can get a transcript of Merlin’s testimony).
But there are plenty of reasons to be concerned for other reasons, too.
HA2’s minor details
First, there are potential concerns raised by the other Russian nuclear engineer, referred to as Human Asset 2, who provided the plan for a 1980s-era Russian fire set to the US in the first place. Scientists at a national lab reverse engineered it, then added in a bunch of flaws that would purportedly prevent the Iranians from being able to use it.
Walter C, who worked at the national lab, described working with the Russian and claimed that — in a series of four meetings at which he apparently did not have much time with the blueprints — he found nothing wrong with them. “It looked credible to him,” Walter C claimed on the stand.
Except Human Asset 2 does appear to have had some concerns. In a cable dated November 25, 1998, Bob S passed on information to the New York and one other office passing on what appears to be Walter C’s message that the two parts Merlin had identified to be missing from the fireset were not supposed to be included.
As we had suspected, the inclusion of certain assemblies on the parts list but not on the schematic was indeed intentional, with the goal of suggesting that the anonymous fireset designed knew that these two assemblies … were essential, but did not know how to make or spec them in any detail.
Please advise [M] of this outcome and suggest that he plan to acknowledge this omission in his eventual presentation to the Iranians if they ask about it.
But the cable went on to reveal that there were other “minor details” that Merlin and HA2 — the original Russian designer of the fireset plans — had raised.
[Mr. G] is still authorized to travel to [the lab] to meet with members of the fireset team to look into some other minor details of the plans which M and [HA2] have noticed or questioned. But per agreement between CP officers and [Mr. C] we will make no changes to the plans and lists until a serious discrepancy arises.
In other words, HA2 may have said, on quick review, that the plans looked credible (at this point, Merlin had examined the plans for what appears to have been an even shorter period of time). But he did, at least, have questions, which is not something Walter C offered up in his testimony.
The Red Team’s 5-month 3-month nuclear project
Then there are other problems in their claims the blueprints wouldn’t be usable. In his testimony, Walter C described how a Red Team of the lab’s nuclear scientists — with 200 years of combined experience! Walter C boasted, as if that were a meaningful stat — tried to find flaws and use the blueprints themselves. While the team only found 25% of the flaws, Walter C claimed, they were able to get it to work in what in testimony (at least according to my notes) was just 5 months.
Only that’s not what the cable said. It said,
After three months of intense effort, and by finding and fixing [some of the] design flaws, the team was able to get a breadboard version of the fire set to work in a laboratory setting.
That’s not the same as using it for a nuclear weapon (as the cable goes onto explain), but at least by my hearing, Walter C misrepresented how long it took the lab to get these flawed blueprints to work.
The mixed assumptions about Russian involvement
Finally, there’s the other way witnesses — especially Walter C — dismissed concerns that the Iranians might be able to use the blueprints. The entire premise for using a Russian blueprint (albeit integrating American and Japanese parts) to distract the Iranians was the Russian brain drain: after the collapse of the Soviet Union, Iran started recruiting newly unpaid scientists. “They had websites looking for areas of WMD concern,” Bob S testified.
Thus, as they tried to build a fireset that the Iranians couldn’t use, they factored in that and “assumed assistance of Russian scientists.”
The export control approval (Exhibit 26) caveated that “in the very remote possibility that the end user can acquire the critical specifications intentionally omitted from the design, and if the user can also acquire the necessary fabrication technologies to successfully fabricate a fully functional device, the end product would then be subject to the above controls.” If the Iranians had access to Russian engineers, as assumed, how could the US be sure they couldn’t use this?
Meanwhile, the enabling technology letter (Exhibit 28) assumes the Iranians had more expertise than claims about their 1950s era nuke program may have let on. The enabling technology approval letter assumed that “User already has a basic understanding of either commercial or nuclear fireset design.”
We don’t have enough data to assess either of these claims (though Walter C’s credibility should be taken about as seriously as his claim that he is “only vaguely” aware of the 2007 NIE finding that Iran had no nuclear weapons program as of 2003). But there seems to be a logical problem with the claims surrounding the plan, that assume both that Iran has access to Russian know-how but that it also doesn’t have access to such know-how.