Richard Lanza, Senior Scientist

Nuclear Engineering, Massachusetts Institute of Technology

February 25, 1998

There are two flights that are crucial to understanding the issue of airport security in the United States: the destruction of Pan Am Flight 103 in 1988 and the crash of TWA Flight 800 in 1996. These two events did not have the same causes – the Pan Am flight was brought down by a bomb and the TWA flight was not. But they were similar reactions after each event: there were public outcries for more money for airport security, and after the TWA flight, Congress did appropriate more money. When the final FAA report came out and said that the TWA problem was likely due to a fuel tank, the budget for transportation security was cut by 3%.

The other important factor is that we think we can solve these problems. In 1990, Congress passed the Aviation Security Improvement Act. This Act puts a high priority on the detection by the United States government of terrorism. It also says we will take every necessary action to detect and respond to terrorist activities. The assumption was that we had the technology to do these things and that the airlines could simply order them off the shelf.

Let me explain airline security very briefly. Looking for a terrorist bomb in airport baggage is like looking for the cure for a very rare disease. We have had very few terrorist bombings and it is very unlikely that the typical baggage inspector will see a bomb in his entire career. Unlike the medical analogy, the methods used by terrorists will change over time. What we really have to ask is what are we going to pay in terms of risk, cost, inconvenience, political change, and changes in our lifestyles to solve this problem? Technology has little to do with these things.

Let’s start off with screening problems. This puts a crunch on technology. First off, this is a rare problem and we want zero false negatives. But it is difficult to determine what your false negative rate is. Another thing that is important is a false positive. The FAA has set a standard of checking one bag every six seconds. The best process has a false positive rate of around 20% -- a bag every 30 seconds is tagged as suspect. Eventually the people running the machines will just ignore suspicious baggage.

Another question is what to look for. A couple of things that current technology tries to look for are shapes and density variations. We also might want to look for the average atomic number and molecular structure. There are others as well.

What are we looking for? The first thing you have to do is figure out what people have in their luggage. The answer is almost anything. The technique that is currently most widely used are metal detectors, but these do not detect explosives. For explosives, the most popular device used is the x-ray.

X-ray systems measure just the x-ray absorption. They look for wires, blasting caps, sticks of dynamite. A variation of x-rays uses a pencil beam that measures the transmission using two different x-ray tube voltages. From this you can get an effective atomic number. If you measure the effective atomic number and density, you find that most explosives cluster around an effective atomic number of 7 or 8, because of the amount of nitrogen. The density varies a far amount, and is less useful. But the atomic number itself gets obscured if things are stacked on top of each other. So, to this you add looking at shapes and then extrapolate. A problem with this system is to find thin sheets of explosives. Another option is to measure scattered x-rays. The current state of the art is an x-ray CT system. They have a pre-scanner to look for suspicious areas and then do a complete three-dimensional reconstruction of a suitcase. The problem is these units costs $1.2 million each. They also have a large misidentification problem and they are designed to look for conventional military-type explosives.

Another option that has been looked at is explosive sniffers. This is attractive because it is a simpler technology and is less expensive. The problem is what to measure. You must determine where to sniff and what to sniff and how smells are transferred. For example, you can transfer smells simply by picking up two pieces of luggage consecutively. One problem is that you need very sensitive sniffers to detect explosives and that you can overload the sniffers rather easily. Then you have to wait until they recover. For example, consider what would happen if chemical compounds related to explosives were secretly added to the floor wax supply at an airport. Then waxing the entire floor at Logan airport could overload the system and shut down the airport because the sniffers were not working.

One option to overcome these problems is to use the systems approach; that is, to use multiple systems on top of each other in a series. The problem, however, is that systems with a high detection probability do not have a high through-put and that the probability of detection is tied to the probability of the least sensitive machine in the system. In other words, a system of machines may actually have a worse probability of detection than single machine systems. There are other problems as well.

Another option is nuclear techniques. You essentially shoot neutrons into the baggage and measure the gamma rays that come out. This allows you to measure oxygen, nitrogen, carbon and chlorine and their ratios to each other. The cost, however, is too great -- $10 to $20 million a copy. Nuclear Quadrupole Resonance is also something that is being looked at, but it cannot detect bombs that are encased in metal.

All the things that have been done to date have concentrated on checked baggage but another problem is cargo. Cargo cannot be scanned with conventional x-ray systems and there is not even a proposed good solution to this. Currently, the FAA claims that since you cannot inspect cargo, you need to rely on the credibility of the shipper.

One final question is "How good does it have to be?" What’s "safe" and what’s "safe enough"? We have set up totally unrealistic standards for what we are expected to do. Right now the reason we have not had anything blow up is not because we have a great security force. It’s because no one has wanted to do it.

Rapporteur: Sharon Weiner

Back to Seminar Schedule, Spring 1998