Eau, Brother!

Electronic noses provide a new sense of the future.

The royal food taster and the proverbial canary in the coal mine: Early methods for detecting toxic substances were usually biological. Over the years, food and environmental risk assessment gradually adopted more refined techniques based on chemistry and physics, but the equipment that’s required tends to be bulky and the processes, slow.

The Nose Chip uses 32 different conducting polymers that respond to various volatile chemicals. The multidimensional pattern created can identify an odor. Cyrano Sciences

This prototype hand-held electronic nose, with an antennalike sniffer port and simple push-button controls, can identify odors in seconds. Cyrano Sciences

Now, researchers are turning back to biology—actually, to electronic mimics of sensory systems. The combination of advanced sensor materials and powerful computer chips holds forth the promise of devices that can sense threats ranging from bacteria in food to the explosives in a land mine. So-called electronic noses are today found mainly in research laboratories, but soon they’ll be sniffing their way into food-processing plants, abandoned battlefields, and maybe even doctors’ offices and home appliances.

A typical electronic nose has two major components: a sensor that detects odors and a set of electronic components that interprets the resulting signals. The sensors incorporate different types of materials, including metal oxides or advanced polymers. When exposed to certain volatile compounds or combinations of them, the sensor materials change size, color, or their resistance to electricity.

By using several different materials in an array of sensors, the electronic nose forms a multidimensional pattern (SN: 1/9/99, p. 31). The device can store this pattern in computer memory and, through exposure to various odors, learn to recognize either a deviation from a good aroma or the presence of a bad one.

Individual components of a sensor need not be large—in fact, dozens can be built onto a computer chip—and they can respond to odors in concentrations well below what the human olfactory system can detect. Furthermore, scientists can design sensors to detect mixtures that even in high concentrations are odorless to people.

Tainted poultry

It’s often obvious to the human nose that food has gone bad, and now researchers are putting electronic noses to that task, too. Judy W. Arnold, a microbiologist in Athens, Ga., and her colleagues have been evaluating the technique at laboratories of the Agricultural Research Service (ARS), the research arm of the U.S. Department of Agriculture. In 1998, they reported that an electronic nose can recognize gases produced by bacteria in tainted poultry.

Since then, Arnold has extended that work and used the electronic nose to identify changes in the aroma of refrigerated chicken that are caused by both the increase of bacteria and the breakdown of fat. Arnold found that the device can detect not just spoilage but also how many days the chicken has been in storage and whether the meat has warmed significantly from a typical refrigeration temperature of 4ºC.

Arnold uses an electronic nose with 12 metal-oxide sensors, each of which responds to a different group of volatile chemicals.

“The poultry industry is very interested in this technology,” says Arnold. “A lot of times, someone at the grocery store will refuse a shipment of poultry because they think it smells funny. But the electronic nose is unbiased and can take the human component out of the picture.”

Other scientists have shown that electronic noses may mercifully reduce the need for human taste testers. A device that can identify changes in the odor of corn oil that occur as it goes rancid, or oxidizes, was described in the December 1999 Journal of the American Oil Chemists’ Society by ARS researchers at Iowa State University in Ames. In subsequent work, the researchers say, they’ve correlated changes in the odor of corn, canola, and soybean oils, as detected by the electronic nose, with disagreeable changes in flavor reported by panels of taste testers.

Nuo Shen, one of the researchers, says that recruiting such panels requires doing mounds of paperwork and overcoming volunteers’ reluctance to sample potentially unsavory food.

“When you tell the volunteers what you need them for, they say, ‘What? You want me to taste rancid corn oil? No thanks, I’m busy!'” Shen says.

A wealth of other research has demonstrated that an electronic nose can be sensitive enough to detect subtle differences among similar agricultural commodities, possibly pointing the way toward its use in the inspection of produce or the verification of a manufacturer’s claims.

For example, an electronic nose using electrically conducting polymer materials in its sensor can tell the  difference between tomatoes that were vine-ripened and those that were picked green. The device can also tell whether tomatoes had been excessively chilled in transit or were bruised inside, even when they showed no outward damage, report Elizabeth A. Baldwin, an ARS research horticulturist in Winter Haven, Fla., and her colleagues.

The same equipment can discriminate between a sample of orange juice made from frozen concentrate and one that had simply been pasteurized, according to Philip E. Shaw and his colleagues, also of the Winter Haven facility.

“This is a sophisticated instrument with a lot of potential,” Shaw says. “The more people use it and the more feedback we give the electronic nose designers, the better it’ll become.”

The electronic nose could be a rapid, reliable, and effective way to inspect fresh and processed meats for quality or even to find out whether choice ground meats have been adulterated with lower-quality cuts, says Arthur M. Spanier, a research chemist at the ARS Meat Science Research Laboratory in Beltsville, Md.

In research presented to an American Meat Science Association conference at Oklahoma State University in Stillwater last June, Spanier showed that the device could discriminate between Serrano hams that have been dry-cured for a full year and less flavorful ones that have been cured for only 7 months. He found that the device could also distinguish among prosciutto, country ham, Virginia ham, and deli ham—a digital connoisseur, indeed. “What the electronic nose does is take an overall look at the bouquet of aromas that comes from the sample, which makes it a great quality-control instrument,” says Spanier.

Moreover, it can perform several hundred preliminary assessments in the time it now takes a chemist to conduct a dozen analyses with traditional methods such as gas chromatography. “Industry loves the quick-and-dirty measurement,” Spanier says.

Beyond the food industry

Electronic noses have promise beyond the food industry. Scientists are using them in other practical applications, such as characterization of environmental air.

With traditional methods of chemical analysis, Jerry L. Hatfield and his colleagues have identified more than 200 different compounds in the air wafting over manure-storage lagoons at large livestock farms. These researchers at the National Soil Tilth Laboratory in Ames, Iowa, have found that 27 of the chemicals are the major contributors to the foul odor that Hatfield calls “eau de lagoon.”

“We can pretty much duplicate the smell of the manure by mixing these 27 chemicals, so we know we’re on the right track,” Hatfield says.

Hatfield is now training an electronic nose to recognize the manure smell. The team plans to attack the odor problem on several fronts, including manipulating the animals’ diet to minimize the offensive chemicals in the manure.

Many crucial questions about manure odor remain. Scientists don’t yet know whether these offensive compounds travel downwind from the manure lagoon as gases or as airborne hitchhikers on dust particles. Nor do researchers know how to obtain good samples of odor at a distance from its source.

“It’s relatively easy to get a water sample from a stream because you know where the stream is,” Hatfield says. “These odors travel in streams, too, but it’s a bit tougher to be sure you’re getting a representative sample from an invisible stream of air.”

Pressing problem

Electronic sniffers are also being asked to address one of an infantryman’s most pressing problems—the detection of land mines. The Defense Advanced Projects Research Agency in Arlington, Va., is now entering the fourth and final year of a project intended to develop an “electronic dog’s nose,” says the program’s manager, Regina E. Dugan.

Land mines threaten civilians as well as soldiers because the mines are often left behind after a battle is over. There are now more than 100 million land mines in the ground in former battle zones in 62 countries worldwide, Dugan says. One in every 236 Cambodians is now an amputee because of land-mine accidents, she adds.

Although bomb-sniffing dogs can detect land mines, Dugan says that the U.S. military uses only metal detectors for the task. The systems can find pieces of metal weighing as little as 0.5 grams and detect even the few small metallic components in modern plastic mines. However, this degree of sensitivity can be a disadvantage because small pieces of metal debris and even some types of soil trigger numerous false alarms. Dugan says that other techniques the Defense Department has tested, such as ground-penetrating radar and infrared imaging, can be fooled by mine-size rocks and other objects.

The agency’s program set out to reduce the number of false alarms by detecting an ingredient in all mines but not other objects—the explosive itself. Several different mine-detection techniques developed under the program have shown promise.

One prototype, developed by researchers at Tufts University School of Medicine in Boston, actually mimics the way a dog sniffs, says neuroscientist John Kauer. The detector moves air back and forth over a sensor that incorporates 32 different polymer materials doped with fluorescent dyes, each of which changes color in response to a specific type of chemical.

The “sniffing” technique prevents strong odors from saturating the sensors, improves the sensitivity of the device, and allows it to discriminate between odors more effectively, Kauer says. The pattern and intensity of colors shown by the sensor indicate the mixture and the concentration of chemicals in the air. The rates at which the colors change indicate the volatility of the chemicals, which helps the scientists identify the particular odor.

A land-mine detector developed by Nomadics takes a different approach, according to Colin Cumming, president and founder of the Stillwater, Okla., company. The Nomadics sensor, which is about the size of a box of cigars, is designed to detect trace amounts of the explosives in most land mines (SN: 3/28/98, p. 202: https://www.sciencenews.org/sn_arc98/3_28_98/bob1.htm).

The sensor uses a single type of fluorescent-doped polymer that glows brightly after exposure to daylight but dims dramatically when even a single molecule of explosive vapor attaches anywhere along the polymer molecule, which was developed at MIT. This property, which Cumming likens to a whole string of Christmas lights going dark when one light goes bad, makes the sensor incredibly sensitive. The Nomadics system can detect concentrations of trinitrotoluene (TNT) as low as 100 parts per quadrillion, which is equivalent to detecting an eyedropper of a chemical dissolved into a volume that could fill 25 Exxon Valdez-size supertankers, asserts Cumming.

In field tests conducted last autumn at Fort Leonard Wood, Mo., the company’s prototype sensor matched the performance of bomb-sniffing dogs that were looking for dummy mines at the same test site. Cumming hopes to have a version of the mine detector ready for limited testing in real mine fields later this year.

Into the kitchen

Research sponsored by the Defense Department has generated a device that’s even closer to rollout and has potential applications that may reach from the battlefield into the home kitchen. With technology developed at the California Institute of Technology in Pasadena, neighboring Cyrano Sciences has designed an electronic device that can identify a variety of odors within seconds, says Saskia Feast, marketing manager for the company.

The device’s sensor, which is built onto a computer chip, incorporates 32 tiny components that swell like sponges and change their resistance to electricity when they’re exposed to various vapors. Feast says these sensors, which the company calls Nose Chips, can be tailored to detect and identify almost any odor. The unique pattern of each odor is analyzed onboard the walkie-talkie-size instrument, which will be introduced at a trade show in mid-March and will sell for just under $8,000.

The devices could be used for purposes as diverse as detecting chemicals in toxic spills, monitoring batch-to-batch consistency in food processing, and helping doctors perform medical diagnoses. Doctors at Children’s Hospital Los Angeles are using a prototype device in a clinical study to see if it can help them diagnose upper-respiratory infections from odors in the breath, Feast adds.

She says the company plans to market Nose Chips to other companies as off-the-shelf components to go into various types of equipment. These sensors, which could sell for $5 to $10 apiece when produced in large numbers, could even be built into home appliances, such as refrigerators and microwaves, where they could detect odors of tainted food or a perfectly cooked dinner.

Asked how she views the future of the electronic nose, Feast says, “That’s easy: I’d like to see them smaller, smarter, and everywhere.”

Taste testing with an electronic tongue

With the same techniques of pattern recognition used in many electronic noses, an electronic tongue can “taste” liquids that people can’t—or shouldn’t—put in their mouths, say researchers at the University of Texas at Austin. Chemist Eric V. Anslyn and his colleagues created a postage-stamp-size chemical-sensing device by drilling a series of small holes in a silicon wafer. They then placed a small polystyrene bead coated with one of several fluorescent materials in each hole. Each bead acts like a taste bud on a tongue.

When the sensor is submersed in a liquid, each bead’s coating reacts to one chemical, and together they yield a distinct pattern of colors. A cameralike device monitors the pattern electronically to identify the chemicals in the mixture.

The original version of this electronic tongue, which the researchers revealed in 1998, used five different fluorescent coatings and therefore could discern only a limited range of tastes.

Anslyn says new versions of the device have a 10-by-10 array of beads that can carry any of dozens of different coatings, enabling the device to analyze more complex mixtures of chemicals.

Anslyn envisions use of the electronic tongue in operations ranging from quality control of food processing to quick, in-office medical diagnostic tests, such as blood and urine analyses.

The electronic tongue’s adaptability is limited only by the types of coatings used on the beads, Anslyn says.

More Stories from Science News on Tech

From the Nature Index

Paid Content