Devices Go Nose to Nose With Bomb-Sniffer Dogs

SAINT-LOUIS, France — Denis Spitzer wants to beat dogs at their own game. At a binational armaments and security research center here in eastern France, Dr. Spitzer and his colleagues are working on a sensor to detect vapors of TNT…

SAINT-LOUIS, France — Denis Spitzer wants to beat dogs at their own game.

At a binational armaments and security research center here in eastern France, Dr. Spitzer and his colleagues are working on a sensor to detect vapors of TNT and other explosives in very faint amounts, as might emanate from a bomb being smuggled through airport security. Using microscopic slivers of silicon covered with forests of even smaller tubes of titanium oxide, they aim to create a device that could supplement, perhaps even supplant, the best mobile bomb detector in the business: the sniffer dog.

But emulating the nose and brain of a trained dog is a formidable task. A bomb-sniffing device must be extremely sensitive, able to develop a signal from a relative handful of molecules. And it must be highly selective, able to distinguish an explosive from the “noise” of other compounds.

While researchers like Dr. Spitzer are making progress — and there are some vapor detectors on the market — when it comes to sensitivity and selectivity, dogs still reign supreme.

“Dogs are awesome,” said Aimee Rose, a product sales director at the sensor manufacturer Flir Systems, which markets a line of explosives detectors called Fido. “They have by far the most developed ability to detect concealed threats,” she said.

But dogs get distracted, cannot work around the clock and require expensive training and handling, Dr. Rose said, so there is a need for instruments.

Flir’s state-of-the-art detectors, including hand-held models that weigh just a few pounds, are used in the military and elsewhere and are generally considered on par with dogs for detecting TNT at extremely low concentrations, a few parts per quadrillion. Even so, Dr. Rose said, “We see our technology as complementary” to dogs.

The devices use a fluorescent polymer technology developed by Timothy M. Swager, a chemist at M.I.T. under whom Dr. Rose studied. Thin films of the polymers emit visible light when exposed to ultraviolet rays, but molecules of TNT stop the fluorescence. A single TNT molecule can quench many thousands of fluorescence reactions, greatly increasing sensitivity.

Dr. Spitzer’s approach is far different, and also far from being commercialized. The slivers of silicon, called microcantilevers, are fixed at one end and made to vibrate, like a diving board. As molecules of explosives are captured by the cantilever, the added mass alters the rate of vibration, which can be measured by a laser or other means.

Microcantilevers are already used in many sensing applications, but for explosives the cantilevers alone are not sensitive enough. Growing hollow cylinders of titanium oxide, called nanotubes, increases the surface area, allowing more molecules to be captured.

Sensitivity is critical because many explosives compounds, including powerful ones like RDX and PETN, are not very volatile — at normal temperatures, very few molecules vaporize. As a result, bomb detection — which became a priority in aviation after a hidden device destroyed a Pan Am jetliner over Lockerbie, Scotland, in 1988, killing 270 people — has focused more on finding explosive particles on surfaces than on detecting molecules in the air.

David Atkinson, chief scientist for explosives detection research at the Pacific Northwest National Laboratory in Richland, Wash., said, “We’ve had a particle-based detection paradigm for the past two decades.” Even today, when a laptop or other object is deemed suspicious after being X-rayed, an agent wipes the surface and puts the wipe in a spectrometer that ionizes any explosive compounds present, allowing them to be quickly identified.

“The key is you’ve got to get that particle,” Dr. Atkinson said. “The whole ability to detect explosives is whether the operator correctly samples.”

Researchers always knew it would be difficult to make an instrument that was as good as or better than a sniffer dog. Paul Waggoner, senior scientist at the Canine Detection Research Institute at Auburn University, said that over the years “the instrument guys” have come to appreciate dogs even more.

“We really are not going to ever be able to surpass the dog in terms of its general ability as a mobile sensing platform,” Dr. Waggoner said.

While the basic mechanisms of olfaction are known, no one knows precisely why dogs are so good at it. Dogs have roughly 30 times as many olfactory cells as people, and the brain region devoted to smell is proportionally much larger — but size doesn’t fully explain it. “The black box that is a dog that we don’t really understand makes a lot of scientists uneasy,” Dr. Waggoner said.

But researchers like Dr. Atkinson have plowed ahead. His team is looking at using the same kind of equipment already in use at airports, but analyzing air rather than particles collected on wipes. The trick, he said, is to increase the efficiency of the equipment so that even a small number of explosives molecules in the air will produce enough ions for detection and identification.

Dr. Atkinson envisages a day when the millimeter-wave imaging booths now used at some airports will be used for chemical detection as well. A passenger will step into the booth, and in the few seconds the imaging equipment does its work, a pump will draw an air sample into a spectrometer for analysis.

At the French-German Research Institute of Saint-Louis, established by treaty after World War II, Dr. Spitzer was inspired less by dogs than by the exquisite olfactory architecture of silkworm moths, described in a scientific article he stumbled upon years ago.

“I read about this kind of silk moth, where the males can recognize females from several kilometers away,” he said. The moth’s antennas were said to be capable of detecting even a single pheromone molecule, and to Dr. Spitzer it seemed that its anatomy, with arrays of bristlelike sensory receptors along its length providing a huge amount of surface area, might have something to do with its sensitivity.

Dr. Spitzer, whose lab spends about a quarter of its time on detection — it also works on highly reactive materials called nanothermites, among other projects — realized that he could increase the surface area of a microcantilever by covering it with nanotubes. And he knew just the person to do this: his wife, Valérie Keller, a researcher in photocatalytic materials at the University of Strasbourg, about 80 miles away.

Dr. Keller and her team had developed a process for growing nanotubes of titanium oxide on various surfaces. Growing them on the silicon of a microcantilever took some doing — one crucial step was to leave enough metallic titanium on the silicon surface so that the nanotubes would not fly off the microcantilever when it vibrated.

The process produces about half a million nanotubes on the microcantilever, which is far smaller than the head of a pin, increasing its surface area by a factor of about 80. That creates far more places for the occasional explosives molecule to find a home.

Dr. Spitzer still has a long way to go. His experiments suggest the microcantilever is capable of detecting TNT at concentrations of less than one part per trillion, which is not yet as good as a dog. The next step, he said, is to determine ways to treat the nanotubes so they are more selective — so the sensor does not also pick up water or other unwanted molecules, which would affect the measurement. It might even be possible to “tune” the nanotubes to capture specific kinds of explosives, he said.

Eventually, Dr. Spitzer will need to test the sensors in simulated airport rooms, under conditions that are as realistic as possible. “When you have real conditions, sometimes you can have surprises,” he said.