VIENNA — As a team of Iranian negotiators parried with the West last week in increasingly heated negotiations over where the country will be forced to dismantle many of its nuclear facilities, one mysterious figure was conspicuously missing.
Mohsen Fakhrizadeh is considered by Western intelligence officials to be the closest thing Iran has to J. Robert Oppenheimer, who guided the Manhattan Project to develop the world’s first nuclear weapon. For more than a decade, he has been identified as the relentless force behind on-again, off-again programs to design a nuclear warhead that could fit atop one of Iran’s long-range missiles — a complex set of technologies that are a critical factor in how long it would take for Iran to build a weapon. As the keeper of Iran’s greatest nuclear secrets, he looms over the talks that he never attends.
“He’s dodging assassins,” one member of the Iranian team said, referring to the “sticky bomb” attacks, widely believed to be the work of Israel, that killed important members of Mr. Fakhrizadeh’s staff as they drove to work. “Wouldn’t you?”
Mr. Fakhrizadeh’s absence underscores a central reality of the increasingly tense negotiations over the Iranian program: If an agreement is reached by July 20, the ostensible deadline, it will be without any real understanding of how close Iran has come to cracking the technologies of building a nuclear warhead. For years, international nuclear inspectors have been told that Mr. Fakhrizadeh is unavailable to talk, and his empire of laboratories and testing grounds is off limits. Now the empire door is being cracked open, just a bit, but investigators say it will take months or years to get any answers.
That means that any agreement among the United States and its five negotiating partners — Britain, France, Germany, China and Russia — within the next few weeks will be shrouded by uncertainty about how long it would take for Iran, if it produced or bought bomb-grade fuel, to make a nuclear weapon. American negotiators seem to be steering away from forcing a full historical accounting from the Iranians before any accord is signed, arguing that excavating the past is less important than assuring Iran does not have the raw material to make a weapon. And the head of the International Atomic Energy Agency, Yukiya Amano, said in an interview last week that no one should expect a complete historical accounting.
“It is not possible to find out everything,” said Mr. Amano, a former Japanese diplomat who is trying, as his predecessor did, to work methodically through a list of a dozen areas that he calls “possible military dimensions” of the Iranian program.
“Some documents have disappeared,” he said. “Some people have already died. In some cases, Iran does not give us access.”
When the subject turned to interviewing Mr. Fakhrizadeh himself, Mr. Amano said that his inspectors have “not yet made a specific request” to see him, but added, “We would like to have access to the sites, the documents and the people — including him.”
It is hardly a new effort. Mr. Amano’s predecessor, Mohamed ElBaradei, who led the I.A.E.A. when it won the Nobel Peace Prize for its efforts, negotiated an ineffective “work plan” with Iran to resolve the issues in 2007; it was never implemented.
By then, the C.I.A. already had a sizable team of experts to study Mr. Fakhrizadeh and his network of government facilities and university laboratories.
In 2004, with the help of allies, they obtained a laptop, quickly given the wry moniker “The Laptop of Death,” that contained documents slipped out of the country by an Iranian technician that contained some of the evidence that Iran was trying to design a weapon.
When the data was shared with the I.A.E.A., the Iranians and others, said it was filled with fabrications. A decade later, many of those documents remain at the core of the I.A.E.A’s inquiry, and Mr. Amano waved aside the claim that they were fictitious.
“That information is included,” he said of the disputed evidence stripped off the hard drive of the laptop so many years ago. “But we spent long hours checking it, and determined the information is broadly credible” because it was consistent with other data, independently gathered by the agency itself. “We can’t say 100 percent, but it is not nonsense information.”
So far Mr. Amano has received responses from Iran on only one of the dozen technologies on his list, and those answers have prompted more questions. He seems unfazed by the thought that his investigation will extend far beyond the deadline for a final agreement with Iran, one that will be made amid considerable uncertainty about how much more work the country would have to complete before it was ready to convert bomb-grade fuel — if it had it — into a weapon.
But gaining any certainty about how much progress the Iranians have made requires getting inside the laboratories, looking at evidence of suspected experiments and digging into reports that Iranians reportedly wrote memorializing their accomplishments.
“You don’t need to see every nut and bolt,” said Olli Heinonen, the former chief inspector of the I.A.E.A., and now at Harvard. “But you are taking a heck of a risk if you don’t establish a baseline of how far they went,” because it would be far more difficult to understand Iran’s timelines to a weapon.
Getting inside is particularly important because American intelligence agencies, stung by their mistakes in Iraq, have changed their views about the pace and political limits of Mr. Fakhrizadeh’s work.
In late 2007, President George W. Bush released an intelligence assessment that Mr. Fakhrizadeh had been told by Iran’s leaders to stand down at the end of 2003.
In subsequent years, some of those efforts, intelligence agencies have since concluded, were resumed in different forms, and under different organizations, but never with the kind of focused effort believed to exist through 2003.
“We probably have to assume that Iran already has a pretty good understanding of how to make a nuclear explosive design work,” said Robert Einhorn, a nuclear expert who worked on the issue until leaving the State Department last year. But access, he said, has a lot to do with monitoring whatever agreement is reached, “because we would have to keep track of the weapons scientists, and know what they are working on.”
Obama administration officials say they have no illusions that they will get visibility into many of Iran’s most heavily protected sites, even if a deal is reached in the next month. That will leave verification of the accord reliant on the American intelligence community’s ability to track covert nuclear activity, a record that is littered with failures.
Iraq’s progress was famously, and wildly, overestimated 11 years ago, helping create the rationale for war. But in the case of Russia in the 1940s, China in the 1960s, and later Pakistan and India, the American intelligence experts often underestimated how long it would take to make a bomb.
As one senior intelligence official said this year, “There’s the obvious fear of repeating Iraq, but anyone with a long memory fears making the opposite mistake, too.”