Phantom study suggests airport x-ray body scans could miss bombs

November 18, 2010

A phantom study suggests whole-body x-ray scanners are not effective. Backscatter scanners unlikely to detect substantial explosive amounts on the body.

Add to the safety and privacy questions about airport whole-body x-ray scanners one more: are they really effective?

A phantom study conducted using a Monte Carlo simulation suggests otherwise and that a substantial amount of explosive could be hidden on a terrorist’s body and escape detection by today’s backscatter x-ray scanners.

Some background: While x-ray backscatter units continue to be deployed at U.S. and foreign airports (“Senators seek more study of whole-body x-ray airport scanner safety), concerns remain unanswered by those with a stake in the deployment of the technology. In a previous note (Radiation risks: Are airport body x-ray scanners ‘a great public health experiment’? ) we briefly argued that contrary to assertions by the manufacturers, government, and assorted apologists, there are issues with penetration, dose and, separately, privacy.

Finally, the justification for all this: The flying public’s safety. But that claim may be questioned as well. We have modeled the machine’s performance using publicly available information on the units and a Monte Carlo program developed by investigators for CERN, the European Council for Nuclear Research. We concluded that only a very poorly informed malefactor would be foiled by these machines. This conclusion has been expressed by others familiar with the technology. (Airport body scanners would be "unlikely" to detect many of the explosive devices used by terrorist groups, a Tory MP has warned; Government Accountability Office says airport body scanners may not have thwarted Christmas Day bombing; Full-Body Scanners Offer a Sneak Peek at TSA Bumbling)

Figure 1 (black and white), shows x-ray backscatter images, on the left from a 125-kVp x-ray source, on the right from a 50-kVp source. The top images are from the Monte Carlo simulation. Depending on assumptions about geometry and detector efficiency in the backscatter unit, the single-view entrance exposure is from 10 microrem (best case) to 70 microrem (more realistic case). In the simulation using the Geant4 human phantom, the subject is carrying 320 grams of PETN plastic explosive. For context, the “shoe bomber” was carrying 40 grams of PETN, and the “Christmas bomber” was carrying 80 grams of PETN. Both amounts were considered sufficient to blow a hole in an aircraft’s skin. The amounts simulated here are four to eight times larger.

The bottom images are provided on the web and for government documents by the manufacturers. Notice at left how conveniently the subject has tied a bag of drugs to the outside of his shoulder and on his sides at the waist, and how he’s placed a gun against a background of tissue. Looking at this image, one could reasonably conclude that the placement of drugs and gun should have been reversed. On the image at right, the low-Z contraband (it does not matter whether it is drugs or explosive or even the same material as the underlying tissue) is packed in a hard edged “brick.” It is detected because of a well-understood hard-edge effect that these machines exhibit.



Figure 2 (color) shows the simulated “bomber.” He is carrying a small (35-ml) water bottle against his body, shown in blue, which is easily seen in the simulation. He is also carrying a 10 x 1-cm piece of iron taped to his abdomen (red in the figure), also easily seen. Finally, he has taped to his abdomen a tapered pancake of PETN explosive, 20-cm in diameter and 1-cm thick at the center, tapering to the edge of the pancake (orange in the color image). The composition of the pancake is approximately 320 grams of PETN explosive, and shows in the simulated images as a slight elevation of intensity, easily confused with normal anatomy (see the images below the simulations). A side image of the phantom (Figure 3) makes the point clearly: All we have done is added a bit of “belly” and exposure. A 40-g pancake would be just 1.25 mm thick at the center and invisible in the image.


Leon Kaufman is professor of physics, emeritus, at the University of California, San Francisco. Joseph W. Carlson is a former professor of physics at the same institution. They have collaborated on many articles and patents.