Schneier on Security
A blog covering security and security technology.
« PDF Redacting Failure |
| Lessons of the ChoicePoint Theft »
May 4, 2005
Detecting Nuclear Material in Transport
One of the few good things that's coming out of the U.S. terrorism policy is some interesting scientific research. This paper discusses detecting nuclear material in transport.
The authors believe that fixed detectors -- for example, at ports -- simply won't work. Terrorists are more likely to use highly enriched uranium (HEU), which is harder to detect, than plutonium. This difficulty of detection is more based on its natural rate of reactivity than on some technological hurdle. "The gamma rays and neutrons useful for detecting shielded HEU permit detection only at short distances (2-4 feet or less) and require that there be sufficient time to count a sufficient number of particles (several minutes to hours)."
The authors conclude that the only way to reliably detect shielded HEU is to build detectors into the transport vehicles. These detectors could take hours to record any radioactivity.
Posted on May 4, 2005 at 7:48 AM
• 20 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
1. It is most likely that a nuclear-enabled -terrorist would be smart enough to not enter/exit through known ports.
2. It is most likely that any transports that detect possible nuclear material would be a false positive. (e.g. http://www.schneier.com/blog/archives/2005/03/...
Does this count as deperimeterization?
Apart from the obvious points stated above, there are better ways of smuggling small atomic weapons than putting them on a boat for several months.
I believe it was Idi Amin who once had a plan to smuggle atomic weapons into his embassies throughout the world through diplomatic channels and then hold the world hostage.
What? These guys are advocating a requirement that every new car sold in the US have a government-mandated snoop-o-matic built into it. Think of all the cool sensors we could put in that package: speed limits, marijuana smoke, vibrations from sexual activities, ... The social benefits would vastly exceed the author's modest goal of . . . of . . . of forcing terrorists to fly their nukes in like loads of cocaine.
And whoever designed Figure 1 loses 99 credibility points: curves showing "Likelihood of nuclear terrorists succeeding" as an *increasing* function of the cost of transporting nuclear contraband?!
I wonder if it wouldn't be just easier for bad guys to enrich the uranium and assemble the device in the US than to smuggle it in. The enrichment process, though delicate, isn't exactly high-tech and is fairly concealable. And uranium's not exactly a rare commodity.
It seems like the devious-foreigners/mad-scientist narrative that surrounds nuclear terrorism inclines us to look in the wrong direction. The equipment needed for enrichment is noticeable in lower-tech and less industrialized countries but might well go unnoticed in the US, the border crossings that have foiled many bomb plots could be avoided, and there could be many fewer people involved in the plot.
Surely anyone who wished to smuggle uranium into the United States would simply hide it inside a giant bale of marijuana; they are imported without detection all the time.
Ever wonder where hospitals get their radioactive tracers? I used to bring deliveries to radiology, after picking them at the bus station. If detectors are placed all around, then every time one of these packages goes by, another hour or so will be wasted determining that they aren't nuclear weapons.
Well, I'm glad to see that someone has finally stated the obvious, that fixed-position detectors won't work for basic physics reasons. But this proposal is really no better, since it relies on the "clients" (i.e. the individual vehicles) to authenticate themselves. What's to stop someone from simply mounting a few inches of lead shielding over their car's radiation detector? Or any other means of circumvention?
(This quite apart from the broader policy issue of requiring that every vehicle in the country have a uniquely identifiable sensor that reports on vehicle contents to every checkpoint it passes.)
More evidence that preemptive wars are a necessary component in the fight against terrorism. We wouldn't have to worry about terrorists smuggling HEU into this country if neither the terrorists or the HEU exist in the first place.
Just a thought,
I remember reading somewhere that the boarder between Canada and the US was not well gaurded with something like one official for every 10Km in the mid 1990's.
Having been on holiday in that part of the world I noticed that the terain is such that it is fairly easy to merge in with the landscape.
So all I do is get my nuke from an ex soviet republic and work out how to get it into Alaska or Canada, then south.
Then comes the really hard part trying to make the dam thing go off. I know the Russian wepons had less fail safes than the US ones but they still had a some.
I suspect that most terorists, do not actually want to make a conventional nucular device, it's not that easy to get a sufficient yield without a lot of knowledge and engineering.
Even though the US released a very large amount of information during the 1970s on the design of Nukes, they kept the information about the initiator (golf ball) secret and it's this that is the major factor in determining the yield.
The clasic argument of a suitcase size "gun style" device is often quoted by pundits (even they know that making explosive lenses for a compresion style device is extreamly dificult).
However if you read up about the "tickling the dragons tail" experiment you will come to the conclusion that although in therory it's easy in practice it's very unlikly likley to go off as it requires critical timing on the three sets of explosives.
If it was easy then just about any nation that got hold of 10 tones or so of uranium could make a device, the fact that very few have is testiment to the problems involved.
If you want to have an effect with radioactive material I suspect the easiest thing to do is wrap is around an explosive charge to make a dirty bomb. The radioactive material for this can be found in quite a few hospitals and construction yards.
Chung Leong -- You're making the assumption that preemptive wars get rid of terrorists and HEU.
It seems to me that evidence suggests the opposite: wars, instability, and failed diplomacy inspire terrorists, risk nuclear materials slipping into the black market where they're hard to track, and incite nations to launch or ramp up nuclear programs.
A strange conclusion actually. What it says to me is that "Active detectors are what is necessary", eg some of these funky proposed detectors which use cosmic rays or a neutron source to bombard the target and see what happens, rather than passive "wait for radiation" detectors.
While I think the article makes some interesting points I am not completely convinced.
First of all behind the physics they use to argue that detection is a fundamental limitation of physics is the assumption that the means of detection is gamma rays or neutrons. What about the possible detection of neutrinos which will pass easily through the shielding and are often released in nuclear decay. Of course there may be an argument why we could never make such detectors of a reasonable size but the point is I am not sure they have really examined all possible means of improving detection.
Secondly, as someone else pointed out this seems to be a strong argument for using some type of active detection technology. It might very well be possible to detect radioactive material with neutrino emiters which could even be used on people without danger.
Moreover, it isn't necessarily true that the only way to detect such material is by its radioactivity. One could also use some type of detector to look for extremely dense pieces of matter like uranium or even lots of heavy shielding. One might even be able to make such a detector that works on some sort of harmless radiation so it could be used in areas with people. Of course one might get some false positives but it could be used to select out things for more detailed examination.
In short their limitations on detection make a great many assumptions which simply may not apply. Still, I tend to think that trying to prevent nuclear weapons via detectors (at fixed points or mobile) is ultimately ineffective. The fact that drug runners can deliver tons of cocaine and heroin via fast boats or otherwise avoiding possible points of detection suggests terrorists can do it just as easily.
In fact if we are talking about a real nuclear bomb US based detectors are completely useless. All one needs to do is bring the boat carrying the bomb into the harbor of a major city and detonate it without ever going through screening. On the other hand trying to close off the borders to small quantities of material used for dirty bombs sounds as hard as stopping heroin trafficing.
In short I am totally unconvinced that the same money which would go to detectors might not be better spent on prevention, or CIA operations or securing nuclear material
The problem with "detection" of any object (not just radioactive material) has two aspects energy and bandwidth.
The object you wish to measure will either emmit energy (from it's own internal sources in a charecteristic way that enables identification).
Or if you push energy at the object it will partialy refelect or transmit it (again this energy from the object will have a charecteristic that enables identification of the object)
Any form of energy is subject to the laws of physics and it's detectable range is related to the energy emmited from the object in the direction of the sensor used to measure it.
I am assuming that the sensor to be used is perfect in that it has no limitations imposed by its design or it's physical realisation. As you will appriciate it is not posible to build such a sensor but you can get quite close if you throw enogh resorces at it ;)
All sensors even perfect ones have a fundemental limitation which is thermal noise, it provides a (bandwidth dependent) floor below which a measurement is not discernable from a random event.
Likewise any charecteristic of the energy (ie it's signiture) from the object needs a given bandwidth for the information to be meaningfull from a series of random events.
For a required amount of information the inverse of bandwidth is time (ie the lower the bandwidth the longer it takes to get the information).
There is a trade off of bandwidth against noise floor, the greater the bandwidth the higher the noise floor.
This means that even a perfect sensor is constrained by what it can measure in a given time of an object at any range.
If you think about it the person trying to hide an object from detection has the following choices if they know about the sensor and it's location,
1, Stay out of range of the sensor
2, Sheild the object from the sensor
3, Modify the objects signiture
4, Move the object past the sensor in a way that does not alow time for detection
Any of these stratagies and most combinations of them will work, so in short all detectors can be bypassed by a knowledgable apponent.
So it is never going to be possible even with a perfect sensor to detect or identify an object with any degree of certainty.
The above is the real reason your last point about the relocation of reasources to more effective avenues of enquiry is valid.
Scenario 1 - truck with a nuke and a working sensor goes past one of the
'checkpoints'. Now presumably these aren't manned and don't have barriers
(as that would be too expensive/restrict traffic flow). So now the powers that
be know there is a potential nuke somewhere past the checkpoint, they might
also have a picture of the truck. So now all they have to do is find it, before
it gets too far into city (not all that likely), and apprehend it in such a
way as to not let the guy inside have a chance to hit the button (again not all
that likely). Unless you have checkpoints _way_ out (which would be impractical)
all you really buying is a few minutes warning.
Scenario 2 - truck with nuke and destroyed/disabled sensor, the guys monitoring
the checkpoints see a bunch of these every day, no big deal - it's normally
a power problem with the truck, send out a patrol car to have a chat with the driver
when he stops. BOOM
Scenario 3 - truck with nuke and hacked sensor that always reports OK - well,
Thanks for your comments Clive. Yes, I do realize that there will always be a trade off between the amount of time you need to detect the object and the bandwidth needed to do the detection and since the bandwidth will be inversely proportional to distance (well probably inversely proportional to the square of distance) one can't hope to have a detector that works at arbitrary ranges in a fixed amount of time.
However, I understood the question at hand to be whether we could build detectors that could work reasonably fast at a reasonable distance. In particular I am not convinced we could not build detectors able to screen every container and truck entering through a controlled border crossing. Sure shielding or speed will prevent this detection but there is a hard limit on the amount of shielding you can put in a truck or container and the screening agency can prevent trucks or containers from zipping through at 60mph.
If the question is whether we could literaly blanket the country or just the borders with detection equitment then I do tend to agree with you that this is probably unworkable. Not because of the fact that there is this tradeoff (if one needed to shield your nuclear material in 100feet of lead and move it at 150mph this would be a major impediment) but because of the particular numbers in this tradeoff (I find it unlikely we would be able to make detectors that worked quickly at large ranges).
In short the goal of any detection scheme is not to make it physically impossible to sneak contraband past but rather to raise the price of doing so (both in technical expertise and likelihood of being caught) to put it beyond the reach of those who would likely violate the law or make it more expensive than other avenues of attack. So it isn't the pure fact that we can't make a perfect detector which is problematic but rather just the expense and difficulty of making one good enough.
Hmm, why would enriched uranium be more difficult to detect than plutonium? AFAIK, plutonium has a half life of about 25000 years - so it hardly radiates at all - it is almost as inert as steel. In fact, coal probably radiates more than a chunk of plutonium. Am I wrong?
There are about 1700 tons of privately owned plutonium already inside the USA. More is being made every single day because our nuclear plant designs were intended to manufacture plutonium for use by the military. Back in 1970, the feds stopped purchasing privately owned plutonium, so the stuff keeps building up in private stockpiles. Some of it is useless for making bombs (even numbered isotopes), and some of it is very useful (odd numbered isotopes).
There isn't much need to smuggle something into the country when it is already here. But that bit of logic would be wasted on the panic seekers.
Chung's comments could be madlibbed into:
"We wouldn't have to worry about aliens smuggling butt-probes into this country if neither the aliens or the probes exist in the first place."
add , spread to everyone
firstname.lastname@example.org to every one
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.