The Berkeley Science Review: Articles

Bombs Away
Hand-held sensors to detect TNT (view PDF)
by Neelang Parghi and Rachel Bernstein

Even after all his scheming, Wile E. Coyote never did manage to catch that Road Runner. But at least Coyote's penchant for TNT, or trinitrotoluene, wasn't misguided. TNT is one of today's most commonly used explosives, valued for its stability and relative ease of use. These characteristics also make it particularly appealing for terrorists and others who might like to use it for destructive purposes. Unfortunately, current bomb detection devices have a high failure rate, often forcing security personnel to search manually for potential explosives. The inefficiency of this approach prompted UC Berkeley mechanical engineering professor Arun Majumdar and bioengineering professor Seung-Wuk Lee to explore new methods for automatically detecting TNT.

The team, which includes graduate students Justyn Jaworski, Digvijay Raorane, and Jin Huh, realized that they could use biology to help solve this problem—proteins, to be exact. Proteins carry out myriad functions in all living organisms, and many of these activities depend on specific recognition of other proteins, nucleic acids, or small molecules. Majumdar and Lee thought they could hijack this natural process to detect explosives as well. "It's like a lock and key," explains Majumdar. "The shape and chemistry of the receptor must match that of the thing you want to detect." However, finding the right "key" for explosive material isn't as easy as going to a locksmith, he says. "It's very hard to machine a key for a molecule, so you try out hundreds of billions of them and see which fit."

To find the best TNT-binding proteins, the team used a technique called phage display. Phages, or viruses that infect bacteria, are made up of a protein coat that encloses the genetic material of the virus. Researchers can manipulate the genetic material to create a vast "library" of viruses that display on their surface any one of a large number of random short peptides, in addition to their natural protein coat. These peptides act like small proteins, taking on a specific shape determined by their sequence of amino acids. This unique shape determines which other molecules the peptide is able to interact with. By using a large peptide library, it is often possible to find a sequence that will bind selectively to a given molecule—in this case, TNT.

To find a TNT receptor, the team doused a solid piece of TNT with a viral sample containing several billion different peptide "keys." The TNT was then washed with a detergent to rinse away any weakly bound viruses, leaving behind just those that held fast to the explosive. Repeating this process with harsher detergents narrowed the field of candidates. The team then tested the peptides that bound well to solid TNT for their ability to bind TNT in solution to ensure that they had identified the strongest, most robust of the lot. "We went from a thousand binding sequences in the first generation to ten in the second, then down to two or three, then down to the single best that detects TNT with a high degree of precision and accuracy," says Lee.

However, for security personnel seeking out hidden explosives, a detector that binds loose airborne TNT particles would be much more valuable than one that must interact with solid material. While TNT has a very low vapor pressure, meaning that few of its molecules escape into the air, this is not the case for dinitrotoluene (DNT), a degradation product of TNT. In other words, any solid TNT is likely surrounded by gaseous DNT that can then be targeted for detection. Following the same wash, rinse, repeat protocol used for the TNT receptor, the group found an effective receptor capable of binding DNT in solution and in the air, even at low atmospheric concentrations amidst all kinds of other potentially sticky molecules.

Finding an appropriate receptor was just the first step of the process. The team still needed to transform the receptor binding event into a visible signal. They have created two devices, each no bigger than a human hand, to accomplish this task. One prototype is composed of a receptor-coated, drumhead-like membrane that is less than one millimeter thick, while the other model places the receptors on tiny metal cantilevers. When the DNT particles in the air bind to the receptor peptides on either the membrane or the cantilever, the devices convert the interaction to an electrical signal that is registered on a computer. Lee estimates that a complete working prototype for a handheld device will be ready in the next three to five years. So while the Road Runner was able to avoid the Coyote's ambushes with speed and cartoon physics, in the real world we may soon have a much better method of detecting TNT.

Neelang Parghi is a science writer living in Berkeley, and Rachel Bernstein is a graduate student in chemistry.

Comments on this article? Drop us a line at with 'letter to the editor' in the subject!