Stardate 20040212.1624

(On Screen): Is the proliferation genie out of the bottle? Wretchard thinks so. I'm not by any stretch happy with the current state of affairs but I don't think it's anything like as dire as Wretchard seems to.

Wretchard has made a number of posts on this subject recently: one, two, three, four. Mostly these seem to have been inspired by the recent revelation of an international collaboration amongst many nations to develop nuclear weapons. I have not seen any complete list of the nations thought to be involved, but at the very least it is known to have involved Pakistan and Libya, and it is virtually certain that North Korea, Malaysia and Iran participated.

At this point I am feeling quite confident that Qaddafi is not playing any games and has genuinely given up his ambition to develop nuclear weapons. There's good reason to believe that Libya's intelligence agency is actively cooperating with the US and UK and revealing everything it has on other nations which were involved. If so, the effort has suffered a calamitous compromise.

Pakistan already had nukes, and its top experts had been providing technical advice to the others. That's now been exposed, but not totally resolved. (Apparently one reason the US is soft-pedaling its reaction is the hope that the Pakistani scientists involved may be able to provide information about the current state of NK's development effort.)

In this post, Wretchard prints a letter from a reader which makes reference to Wretchard's article from last September that describes "Three Conjectures" (which I discovered and wrote about late in November.)

Wretchard's reader lays out a scenario whereby dozens of working nukes would be smuggled into the US before any of them were actually used. Wretchard largely accepts the assumptions behind that scenario and develops it further.

But both of them have made what I think is a fundamental mistake. They both seem to assume that the hardest part of the process of developing nuclear weapons is coming up with a design. Wretchard's reader SR begins:

It's SR again. Ok, I'll be honest, The Three Conjectures has been nagging me for months. So why am I writing you now? I realized that there was an implicit assumption made in the 2nd conjecture. The explicit assumption in 2nd conjecture was "Suppose Pakistan or North Korea engineered a reliable plutonium weapon that could be built to one-point safety in any machine shop with a minimum of skill, giving Islamic terrorists the means to repeatedly attack America indefinitely". The implicit assumption was that the Islamists would start with a single iteration, i.e. attack cities one or two at a time. There is no reason to assume that they would do so.

Actually, from what I've read it is not very hard for someone with the right education and access to certain open sources of technical information to develop a working design for a nuclear weapon. In one notorious case, an American Ph.D candidate in nuclear physics actually did it in his thesis.

People in the know tend to be cagey about this kind of stuff. I've got one long-term reader who knows a very great deal about that kind of thing. We've been exchanging email for nearly as long as this site has existed and I always enjoy receiving letters from him. A couple of months ago I sent him a letter asking him a few technical questions about why there were such radical differences in the designs of weapons designed to use U²³⁵ versus those designed to use Plutonium. He didn't respond to it, and I didn't press, taking his silence as an unwillingness to discuss such things.

The limiting factor has always been access to adequate quantities of fissionables, and international non-proliferation efforts have always primarily concentrated on that.

The process of producing large quantities of purified plutonium and/or U²³⁵ enriched enough to be fissionable is slow, requires large amounts of expensive and customized equipment, is expensive, and consumes enormous amounts of power. At least some of the facilities involved in such efforts will be physically large. As a practical matter, it can only be done by a government, or with both knowledge and approval of a government.

The two are different, but neither is easy. U²³⁵ only makes up about 6/1000ths of naturally occurring uranium, but it has to be enriched to well above 95% in order to be used in a weapon. Chemically speaking, U²³⁵ and U²³⁸ cannot be differentiated since they're both uranium, so the only way to separate them is by mass, taking advantage of the fact that U²³⁵ weighs very slightly less.

Much of the recent revelations relate to development of gas centrifuges. That approach takes naturally-occurring uranium metal and treats it with hydrofluoric acid to create uranium hexafluoride, which becomes a gas at temperatures above 56° Celsius. The weight difference between U²³⁵F₆ and U²³⁸F₆ is less than 1%, but if it is spun very rapidly in a centrifuge there is a slight tendency for the lighter molecules to rise. That can be skimmed off, yielding a gas which has a somewhat higher concentration of U^235, which can then be put through the process again to enrich it a bit further. If you begin with a couple of metric tons of natural uranium and run your process to maximize yield, then at the end you'll produce enough highly-enriched U²³⁵ for one weapon.

But you need a lot of centrifuges and a lot of space for them and a hell of a lot of power, and it takes a long time because the process is so inefficient. If you have more uranium to begin with and are willing to accept less total yield, it can be done a bit faster, but if you want to maximize the yield, it's extremely slow.

Plutonium can only be found in trace quantities in nature. (For a long time it was thought that it didn't appear naturally at all.) The only reasonable source of it is from processing fuel rods after they have been used in fission reactors. The uranium in the fuel rods used in such reactors is also enriched, but usually only to a concentration of a couple of percent of U²³⁵.

The vast majority of the uranium in the fuel rods is U²³⁸, which is not fissionable. When the small fraction of the fuel which is U²³⁵ fissions during normal reactor operations, it tends to release neutrons and many of those will be captured by neighboring U²³⁸ atoms, converting them to U²³⁹.

U²³⁹ β- decays with a half-life of 23.5 minutes, yielding Neptunium²³⁹. That, in turn, also β- decays with a half-life of 2.35 days, yielding Plutonium²³⁹, which is fissionable.

The good news is that the yield per initial kilogram is higher than with U²³⁵ enrichment, because the seed stock is plentiful U²³⁸ rather than rare U²³⁵ (though only a small percentage of the U²³⁸ converts). Even better, plutonium can be separated out of the fuel rods and purified using chemical reagents, since it is a different element and chemically different. Note that it is not easy, even so, but it is not nearly as bad as mucking around with gas centrifuges.

The bad news is that the process of converting U²³⁸ into Pu²³⁹ only happens when the fuel rods spend months in a real operating reactor, and all the reactors available to members of this nuclear underground do require slightly-enriched uranium for fuel. And for the most part, both the fuel and the spent fuel-rods have tended to be monitored (albeit with widely varying degrees of accuracy and scrutiny). So it's difficult to do this without being noticed.

(I just thought I'd mention in passing that I am deliberately skipping a lot of details and deliberately making a few statements which are not quite true. There are also other more obscure fissionables which can be produced using other processes, but they are not practical for the members of the nuclear underground. So DWL!)

That's why I don't really agree with Wretchard when he says this:

There is no reason why a 30 city first strike is inconceivable, or even difficult as this 40-year old design shows. Nuclear weapons are cheap. According to the Brookings Institution, the United States built 32,000 warheads since 1945 at a total cost of $409 billion in 1996 dollars -- about $13 million apiece.

"Inconceivable" is one of those terms which obscures more than it reveals. If one interprets it to mean "impossible", then it's true that such a strike is not impossible. it does not, after all, violate the laws of physics or involve logical contradiction. But that doesn't imply that the chance of it happening is significant, and in fact the chance of such a plot right now is so small as to be ignorable. There is no indication that the members of the nuclear underground have advanced far enough and operated long enough to produce enough fissionables for such an attack, and there are other reasons why it is implausible.

Wretchard's references to US costs is not really relevant, because it's an apples-and-oranges comparison. This is very definitely a case where economy of scale operates, and the per-warhead cost he calculates is due to the fact that the American program was huge, long-lasting, well-funded and unmolested by nasty neighbors and international inspections agencies, and was subject to significant improvements as time went on and knowledge was gained. We cannot conclude that it would be equally easy and/or cheap for the nations involved in the nuclear underground.

They were trying to do all of this surreptitiously, and secrecy costs a lot. (One of the reasons they seem to have been concentrating on enrichment of U-235 is that it's easier to cover your tracks.)

Now suppose, hypothetically, that these programs were extremely advanced and had actually been operating effectively for several years. Suppose that they had managed to produce enough fissionables for 30 weapons. Does it make sense that they would assign all of it to a highly risky scheme involving smuggling into the US? It seems to me that the nations in question would want most of their production of fissionables to be used to produce weapons for a national arsenal, and indeed that they'd want all of it used that way until said arsenals had been built up. You would not see any diverted into terrorist hands initially.

But would they eventually be willing to give even a small percentage to terrorists? Do those governments really trust such terrorist groups enough to bestow such weapons on them? Wretchard mentions Saudi Arabia's immense oil income, though ignoring the fact that their economic situation has degraded immensely in the last 20 years and that their current income isn't enough to prevent internal unrest. But given that al Qaeda has been making attacks inside Saudi Arabia, then if the Saudi government were a party to this effort, would it be willing to permit one or two weapons-worth of fissionables to get into al Qaeda's hands, facing the risk that al Qaeda might use it against Saudi Arabia instead of against us? It seems highly unlikely.

I haven't seen anything to suggest that Saudi Arabia was part of this (or was not); I'm responding to Wretchard's example. Iran is, however, part of it. Iran's relationship with al Qaeda has always been love-hate, given that al Qaeda is fundamentally extremist Sunni and Iran's ruling theocrats are extremist Shiite, and each thinks of the other as heretics. Iran has given refuge to some al Qaeda members in the aftermath of our operations in Afghanistan, but I see nothing to indicate that they've buried the hatchet and become full allies. I don't believe even Iran would trust al Qaeda enough to give them enough fissionables for even one warhead, since they could not be sure it would not ultimately be used against them or their friends instead of against us.

Note recent revelations of al Qaeda plots to try to stir up a civil war in Iraq between Sunnis and Shiites, by making attacks against Shiites in hopes of provoking a backlash. Would Iran risk the chance that al Qaeda might use such a nuke on the Shiite holy city of Najaf instead of New York? Certainly it would be a lot easier to smuggle such a thing into southern Iraq.

It is said that one should never underestimate an enemy in war, and that's true. But it's nearly as bad a mistake to massively overestimate one, and I believe that Wretchard's reader SR does so when he says this:

Keeping in mind that Al Qaeda is known for two operational principles: the ability to wait for years, if necessary, between attacks in order to position its resources and its use of simultaneous operations. Assuming the leadership learned its lesson in Afghanistan and Iraq, they know U.S. is not a paper tiger unwilling and/or unable to go to war against its allies wishes. Therefore, there is no reason, aside for logistical difficulties, for them not to wait and use a say, 30-40 city simultaneous attack.

I do not think that al Qaeda is quite the patient, well-organized, well-led operation he gives it credit for being. It's true that they waited long periods between attacks against us, but that was mostly because of limited resources and limited opportunities.

But we can't obsess on our own problems while totally ignoring the problems al Qaeda faces. Support is not a sure thing for them; they don't have a stable income from a tax base or a stable source of recruits. In order to keep the money and recruits flowing, they have to produce a high-profile success occasionally to maintain their reputation. And they have not done so.

The larger the plans they engage in, and the longer such plans take to reach fruition, the greater the chance that such plans will be detected before they can be executed. SR says:

Therefore, there is no reason, aside for logistical difficulties, for them not to wait and use a say, 30-40 city simultaneous attack.

Logistical difficulties can't be dismissed as lightly as all that, since I don't see any plausible scenario whereby they could come up with enough fissionables for an attack of such magnitude. But even if they could, the sheer size and complexity of the process of setting up such an attack would make the security risk too great.

The reason they are organized in cells is to minimize the damage when a cell is discovered. By the same token, they are unlikely to engage in an operation this huge because if it is compromised they lose too much all at once. If they had ten working nukes, they'd start ten totally independent plans instead of one grand ten-attack plan.

But that's not the least of their problems. It is not at all clear they actually have the human resources to engage in a plot of this magnitude. It is now known that the 9/11 attack was launched using their A-team; it required most of their best-quality willing martyrs just for that one operation.

Donald Sensing has written quite a lot lately about the problems al Qaeda faces, and how much trouble they are in. They do not in fact have the ability to wait for years before launching their next major attack against us, because they are suffering steady attrition from our counter-operations, and because support for their cause will decline as time goes on.

And as they have begun to concentrate less on killing heathen and more on killing Muslims, in Iraq and in Saudi Arabia and elsewhere, their support will decline further.

It took them years to plan and set up a plot for jet hijacking which was executed on 9/11. 4 jets were hijacked, three reached high-quality targets. There are rumors that there may have been a couple of other teams whose flights were cancelled, though I have never seen any confirmation of that.

But since then, they've been under serious pressure in nearly every way, and that isn't going to let up. They lost their best base of operations; the majority of their top people are gone; much of their income has been cut off; and their reputation is declining rapidly amongst those populations which had been their primary sources of recruits and contributions.

I think I can plausibly argue that if al Qaeda had enough fissionables for a single nuke, they'd be working to use it immediately. They would not wait until they had as many as SR and Wretchard describe.

I cannot say that the 30-city scenario they describe is inconceivable but I do not find it remotely plausible. I don't believe that the resources are there, and I don't believe that it's the kind of plan al Qaeda would make even if the resources were there.

I'm afraid that Wretchard and his reader are engaging in the kind of worst-case fear-mongering that opponents of the invasion of Iraq engaged in last year, what with estimates of tens or even hundreds of thousands of Iraqi civilian deaths and millions of refugees, along with plague, starvation, stubbed toes, acne, and every other possible negative outcome they could imagine. I don't think Wretchard serves his position well by engaging in equal hype, because he actually has a legitimate point beneath this all with which I agree, one which might be ignored by a reader who doesn't take Wretchard's gloom-and-doom scenario seriously.

There's no question that it was extremely fortunate that this multi-nation plan was exposed when it was, even though it's likely that either Iran or North Korea have already managed to create enough fissionables for a few weapons. It would have been far better if it had been exposed earlier, but an additional ten year delay would have been catastrophic.

There is equally no doubt at all that those revelations were a direct consequence of the war the US is prosecuting, most especially in Iraq. It is by no means clear that without our war there would have been a significant chance of the plot being exposed until too late.

There is no doubt at all that previous nuclear non-proliferation programs and policies are now exposed as utter failures, which will have to be replaced with something more effective.

But it is not at all clear that the nuclear proliferation genie can no longer be bottled. That will become more clear in the next few months, but I don't think the overall situation is anything like as bleak as Wretchard and SR portray it as being.

Update: Wretchard responds.

Update 20040213: TMLutas comments.

Update 20040214: I'd like to thank all fifty people who wrote to tell me what the difference was between Uranium and Plutonium that affected bomb design. Actually, I already knew the difference. (If you look carefully, you'll find that I said that I did not know two months ago, but said nothing about current knowledge, and did not ask for current help.)

Update: I guess that was too subtle. Let's try being more blunt: Please stop sending me information about the difference between uranium and plutonium. (My mailbox is looking like it did last July.)