Home About My Books Résumé Chronological World of Wonders

Some notes on x-ray backscatter imaging

Created/Modified: 2010-11-17/2010-11-17

This note has two parts: one on why x-ray backscatter imagers aren't safe, and the other on why they don't work.

They Aren't Safe

There's a lot of stuff in the news lately about the deployment of x-ray backscatter imagers in American airports. To be honest, being seen naked doesn't bother me, and I find the whole "nudity taboo" thing is one of the weirder aspects of human behaviour. But unlike the organs of the security-industrial complex, I respect people's right to their neuroses, particularly as I bloody well expect them to respect mine.

That said, how safe are these machines? And how reasonable are the standards that are being applied to them?

This open letter from a group of scientists raising concerns regarding the safety of these devices and questioning the appropriateness of the testing that has been done on them to date is entirely to the point, in my view as a radiation transport physicist who is intimately familiar with the processes by which x-rays deposit energy in tissue.

Ask any radiation oncologist what the big problems in treating cancer with radiation are, and at some point they will say the words, "skin dose". The reasons why skin dose is a problem for the kind of high-energy x-rays used in cancer treatment are a little different from the problem of skin dose with low energy x-rays used in back-scatter imagers, but the problem remains.

They say a picture is worth a thousand words, and I usually leave the art to Hilary for what are about to become obvious reasons, but in this case I've given it a shot:

Skin Dose

In the picture we've got x-rays incident on the skin from the left. The skin is the thin layer between the two black verticals. To the right of skin is the rest of the body, labelled "meat", because that's what we're made out of. Superimposed on this cartoon of the body is a graph showing x-ray energy deposited as a function of depth, which is roughly exponential. As the blue curve was hand-drawn, perhaps that should be "very roughly".

The blue horizontal lines give the average dose over the whole curve, and the average dose to the thin layer of skin that is up-close-and-personal with the incoming x-ray beam.

The difference between the average dose--which is what has been measured--and the skin dose--which is what people without a pecuniary interest in the matter are concerned with--I've labelled with "SC", for "Skin Cancer".

I said above that I didn't much care about being seen naked, but I appreciate other people's right to differ on that question.

What I do not appreciate is being exposed to an unquantified risk of skin cancer from a potentially very high skin dose that has not been properly measured or accounted for in the standards-based radiation safety testing of these machines, which is not well-suited for systems that are designed primarily to give high surface doses.

It isn't rocket science to do the job of testing these things properly. I have a good deal of expertise in the niceties of radiation transport in thin-layered systems, and there are some nuances, but given the millions someone is making on these things there ought to be plenty of money available to ensure their basic safety.

As it stands, these machines are a pure act of state-sponsored violence against anyone who is misled into one. I would certainly not allow myself to be scanned by one of these machines, nor allow one of my minor children to be scanned by one.

Millimeter wave imagers have the same issues with privacy and nudity and prudity that x-ray backscatter imagers do, but almost certainly don't pose any major safety issues, although massive public exposure is a hell of a way to perform the experiment.

Based on a back of the envelope calculation with a 50 KVp Bremsstrahlung spectrum and the standard mass absorption coefficients for soft tissue, X-ray backscatter machines will, if widely deployed, almost certainly kill far more people than the terrorists they are supposed to protect us from.

They Don't Work

In 2012 scientist, engineer and activist named Jonathan Corbett who has demonstrated a trivial way of defeating the x-ray backscatter machines I wrote about a while back.

The popular response to Mr Corbett's demonstration has received a stunningly unprofessional, incompetent response from the TSA, which thanks to the Streisand Effect has made the news that x-ray backscatter scanners can be easily defeated much better known than it would have been otherwise.

But it is worth pointing out that the problem was known in the literature sometime ago. You may need an academic account to get access to the linked paper, but the gist is that a couple of people have used GEANT4--one of the standard radiation transport codes, which is known to work well at low energy--to do some simulations of the x-ray backscatter systems, and concluded that they can be easily defeated by a variety of simple techniques, and this fact depends only on the physics of radiation transport, not on any implementation details of the scanners.

So if you are hot to suggested modified technique for using the scanners, you are barking up the wrong tree. You need to talk to God and get him to change the laws of physics to make these things work. Although that still won't make them safe (see above).

The problems with these systems are two-fold, and depend on the most fundamental aspect of how ionizing radiation interacts with matter. Light elements, which have low atomic number (hydrogen, oxygen, carbon, nitrogen... the things that humans and plastic explosives are made of) tend to scatter low-energy x-rays rather than absorb them.

The amount of scattering is not hugely sensitive to density and composition of the low atomic number material, which means--as the above-linked paper from the Journal of Transportation Security points out--that it is really hard to detect plastic explosives that have been "pancaked" on to a person's torso. The only reason they stand out in the TSA faked demonstration images is that they are conveniently brick-shaped, which produces edge effects that are easily detectable. Remove the sharp edges by smoothing the stuff out and you lose the ability to detect it, not as a matter of how the instrument is deployed but as a fundamental consequence of the Compton scattering cross-section as a function of atomic number. The only way around this is to increase the dose and/or increase the beam energy (which increases the dose.)

Secondly, and more germane to Mr Corbett's successful demonstration of how to defeat this crude technology, is that it provides no way to tell why no x-rays are scattered back from a particular region. There are two possibilities: a) there is no scattering material there (air around the body, for example) and b) there is a strongly absorbing material there (the metal case in Mr Corbett's successful circumvention demonstration).

All the detector knows is that it isn't getting any x-rays back from a particular region, and it has no way to distinguish between air and metal: neither return a significant backscatter signal. This is why the background of the backscatter images is black, and why the suggestion of "just display the image on a chequerboard background" misses the point entirely: the black background is not an arbitrary choice, but determined by the basic physics of the apparatus. Any region that returns no scattered x-rays to speak of will appear as black, because the brightness of the image just encodes the flux of scattered x-rays.

So there are at least two well-known, well-documented ways of defeating these unsafe scanners, and yet no one has blown up any American aircraft recently, possibly because the United States has far fewer enemies than the Organs of the State would have one believe.

Finally, there is a point to be made regarding the TSA response to Mr Corbett's demonstration: pointing out all the cases where a technology does work is completely irrelevant to the very important cases where it does not work. Mr Corbett has pointed out that x-ray backscatter scanners fail in certain very important and easily realized cases. Blogger Bob has "responded" by pointing out--amongst other things--that in a variety of completely unrelated cases the technology does work.

This is not useful. It is disingenuous and insulting to the intelligence of the average traveller. It is like saying that a fist-sized hole in the hull of a boat doesn't matter because the hull is perfectly sound everywhere else. True, but hardly relevant to the massive amount of water pouring in.

Anyone who wants to can walk almost anything by the ineffective and unsafe x-ray backscatter scanners by avoiding the cases where they do work and exploiting the well-known and well-documented cases where they do not. Maybe it's time to start talking about the things that will really make Americans safer--like genuinely universal health care, a much smaller military, and less involvement in what "some guy" referred to as "foreign entanglements".

Contact Home World of Wonders
Copyright (C) TJ Radcliffe