The vapor chase: in the quest to build and sell a mechanical nose, experts have created technologies almost as versatile and complex as the human sense of smell—almost – Body Issue – eNose/Cyranose

The vapor chase: in the quest to build and sell a mechanical nose, experts have created technologies almost as versatile and complex as the human sense of smell—almost – Body Issue – eNose/Cyranose – Brief Article

Joshua Tompkins

IN GERMAN NOVELIST PATRICK SUSKIND’S 1985 PERFUME: THE STORY of a Murderer, the main character is a man born without body scent yet with a freakishly acute sense of smell. Jean, Baptiste Grenouille recalls his childhood through odors–“the hostile, steaming vapors of Madame Gaillard’s bedroom, the bone-dry leathery bouquet of her hands … the carrion stench of the Cimetiere des Innocents.” He laments the failure of language to describe the Thousand teeming nuances his nose beholds with each breath. * His own lack of odor causes people to avoid him with dread, as if he cast no shadow or produced no body heat. Having no smell, Suskind seems to be saying, is akin to having no soul. Humans are mysteriously attuned to one another’s gaseous halos. Grenouille cannot bear the stench of humanity and imagines himself as a living alembic; he distills the essences of nature’s most beautiful creations into the greatest fragrance the world has ever known. The resulting potion, less a perfume than simply the ultimate human scent, is so seductive that the scene of his execution for murder soon becomes the wildest orgy since Caligula’s Rome.

Smell, as Grenouille’s exploits suggest, is both the mind’s muse and its hobgoblin. It is also the most underrated sense. When polled, people invariably choose it as the one they would most readily relinquish. The most primordial yet least understood sense, it is an instinctual, bestial spirit guide that helped primitive humans locate food, mates, and predators. The Botocudos people of eastern Brazil and a few aboriginal tribes on the Malay Peninsula still track their prey by smell, but the rest of us possess noses blunted and co-opted by herbal shampoo and scented slot machines and that new-car smell. We notice stimuli only at their extremes: the siren succulence of a hundred barbecues on a summer Sunday in Griffith Park, the netherworldly fetor that has loitered for years at the corner of 6th and Fairfax. We are unaware of the unique odor signature each of us radiates, oblivious to the curious hormonal link between the nose and the sex drive. We find it impossible to describe a fresh smell without making reference to a familiar one. We cannot mentally conjure a scent the way we bring a face or a line of music to mind. In fact, there is no verb in English–or Spanish, or French–that means “to smell good.”

Despite the neglect, researchers in Southern California are competing to deconstruct the muse. While some focus on the mechanisms of odor perception, others have created computerized sensors–known as electronic noses–that are in some cases more sensitive than their human counterparts. Like Suskind’s man, one portable device can even identify the components of a complex odor with a single sniff.

The modern quest for mechanical olfaction began with the rise of the first electronic instruments. In 1920 two European scientists proposed measuring the electrical charge on a spray of scented water but didn’t pursue the matter. Some 40 years later in the United States the first electrochemical detectors were employed, yet no real effort was put forth to make sense of the data. Only in the mid 1980s, with advances in computing and electronics, did the concept of the artificial nose emerge as an intelligent system of sensors configured to pick up a wide variety of chemicals. The term electronic nose soon appeared in literature for the first time.

Caltech’s involvement with the electronic nose began about ten years ago, when chemistry professor Nate Lewis correctly predicted that olfaction was simply a matter of pattern recognition. Humans are capable of distinguishing some 10,000 different odors, with each scent triggering a characteristic set of responses in a person’s nasal receptors, which in turn transmit a unique signal to the brain. Thus, a 1997 Chateau d’Yquem doesn’t smell quite like a bottle from the year before–even to someone who hasn’t previously experienced either.

Lewis realized that ‘duplicating human odor perception required the integration of two systems: an array of sensors that could generate unique patterns for as many smells as possible and a sophisticated computer program, known as a neural network, to learn the patterns the way the brain does. “You just need enough sensors sending enough signals,” he said in 1996, “so that no two things have the same pattern and seem to smell alike.”

More than five years later the number of sensors ranges from 32 to 32,000 or more, and the resulting system replaces neurobiology with polymer chemistry. Each of the e-nose’s sensors is sprayed with a different polymer (a specialized plastic), creating a circuit that conducts very little electricity and provides a measurable level of resistance. When an odorous vapor passes across the sensors, the brief increase in resistance allows a computer connected to the sensors to identify the vapor.

There are now several different editions of the e-nose at Caltech, all of them crammed into a shopworn basement laboratory with room for only two researchers to work at the same time. The biggest nose covers a sizable portion of one wall. Large bottles of organic liquids are racked in neat rows and connected by a network of flexible tubes to the sensor array, which at first glance could be mistaken for a household outlet strip with the cover pried off. Near the ceiling a thick exhaust hose provides ventilation with a touch of metaphor: Dark gray and wrinkled, it looks like an elephant’s trunk poised to take a whiff.

Placed under a whooshing, pale blue fume hood, the nose of researcher Eric Tillman is the size of a footlocker and bears more resemblance to a high school chemistry lab station than to the Nutty Professor-like rig on the other side of the room. The workspace of Tillman, who is somewhat baby-faced with a close-cropped head of hair, includes a pair of pliers, various plastic fittings, a large syringe. He is currently teaching the nose to recognize the organic compounds associated with illnesses such as cirrhosis and breast cancer, for which the e-nose could one day serve as a screening device.

Tillman demonstrates the technology using butylamine, an organic compound that bears a strong suggestion of poor hygiene. When just a tiny amount of the diluted vapor is injected into a bell jar containing the sensors and a magnetic stirrer that circulates the air, the graph on a nearby computer rockets off the top of the screen. The e-nose, in this case, is more sensitive to the odor than the human nose.

There are limitations. The e-nose couldn’t serve as a household carbon monoxide detector because the gas is too volatile for the sensors to detect. Nor could it help a patent thief unravel the recipe for Coca-Cola or Quelques Fleurs. It doesn’t deal with the individual components of a smell, just the overall bouquet. Plus, the e-nose still exhibits certain olfactory blind spots, most notably its inability to detect thiols, a family of sulfurous chemicals that give rotten eggs their familiar odor. Lewis says the nose’s palate might be expanded by further tailoring the material used in each sensor. That way, the customized sensors could enhance the sensitivity to include most of the chemical spectrum.

Those sensors needn’t be as large as the ones in the lab. In his office, which has the greenish feel of excessive fluorescent lighting, Lewis holds up a prototype chip with a gold mosaic grid of circuits packed into an area smaller than a thumbnail. Every cell phone could easily be outfitted with such a chip, he says. As a patient talks to his or her physician, the e-nose could sniff for chemicals indicating disease and transmit the data to the doctor’s computer screen. Mass produced, the chip would be inexpensive. Says Lewis, “It would cost about ten cents.”

Beside Lewis’s desktop computer sits a reminder that both the e-nose and the human version pale in comparison with the ultimate airborne-chemical detection system. On a coffee cup Michelangelo’s God is no longer touching the fingertip of Adam. He’s blessing the nose of a German shepherd.

THE STUDY OF SMELL HAS LONG SUFFERED FROM ECCENtric Connotations and persistent upstaging. More is known about Cision and hearing and how the brain interprets the input from those senses. Aristotle classified odors in but four categories: pungent, acidic, succulent, and astringent. Certain smells earned a reputation in medieval Europe, where “bad odors” such as the stench of an epidemic victim were believed to carry the disease itself. “The smell whereof shall breed a plague in France,” wrote Shakespeare in Henry V, and the nursery rhyme expression “pocket full of posies” refers to the nosegays terrified citizens held to their faces to ward of baleful vapors. Some even believed the bowels of the earth housed a “stench laboratory” that sent mephitic miasmas roiling to the surface to sicken people in great numbers. Immanuel Kant reviled smell as the lowest-ranking sense, arguing that unpleasant odors outnumbered nice ones. In a 1918 essay Alexander Graham Bell urged a quantitative approach to the subject. “We have very many different kinds of smells,” he wrote, “from the odor of violets down to asafetida, but until you can measure their likenesses and differences you can have no science of odor.”

The subjective mechanics of odor perception–that is, why mint smells minty and a rose unlike a daffodil–are still poorly understood. The physiology, however, is clear enough. When a person sniffs a bouquet of roses or a musty box of heirlooms, molecules inhaled through the nose flow through the sinus baffles, where they are warmed, to a pair of dime-sized regions at the top of the nasal cavity called the olfactory epithelia, on which millions of specialized nerve receptors are crammed. There, about 2 percent of the odorant molecules, snared in a mucous layer, activate a few of the receptors, which in turn transmit signals through the bone over the nasal cavity. The signals converge in a small structure at the base of the brain called the olfactory bulb.

Unlike the other four senses, smell bypasses the thalamus–a place as central to mental function as Memphis is to Federal Express, routing all other inbound sensations and outbound motor commands–and heads directly to the olfactory cortex, where the inhaled odor is perceived by the conscious brain for the first time. This neurological hot-wiring might help explain why a scent can evoke memories more poignantly than an old photo or a whistled melody. Marcel Proust, in the introduction to Swann’s Way, experienced an epiphany after soaking a few crumbles of a madeleine in a spoonful of tea; to encounter a long-lost fragrance, he reckoned, was the only true means of traveling in time. “When nothing else subsists from the past, after the people are dead, after the things are broken and scattered … the smell and taste of things remain poised a long time, like souls.”

Another bodily reason exists for the reverie induced by familiar scents. Smell, as any aromatherapist will explain, is the only sense directly connected to the limbic system, the primary seat of emotion, motivation, and recollection. Here the signal reaches a small structure called the hippocampus, which among other tasks processes long-term memory. In Proust’s case the stimulus matched a dormant smell memory of childhood mornings with his Aunt Leonie; the shudder of pleasure he described as a result likely happened when his hypothalamus sent an impulse via neurotransmitters that caused the muscles in his neck and back to spasm.

Unlike most types of human nerve tissue, in which cells are irreplaceable, the receptors in the nasal cavity are routinely renewed. As Professor Lewis might phrase it, the entire sensor array is periodically dumped and a new one wired up. Yet sandalwood and diesel exhaust smell just as they did a few months ago. Newborns don’t lose the ability to recognize their mothers by scent, nor children their siblings. The emotive link between smell and memory is in this way one step demystified: A scent conjures the remembrance of things past by first conjuring the remembrance of the same scent.

THE LICENSE TO CALTECH’S FIRST-generation e-nose technology is held by Cyrano Sciences, headquartered about two miles northeast of campus in a light-industry park off Colorado Boulevard in Pasadena. Since September 2000 the company has sold about 200 units of its Cyranose 320, a handheld device that weighs two pounds and has a small LCD display. “Sales have been slower than we would like,” says president and CEO Steven Sunshine. Most of the buyers are industrial, such as an Alaskan fish processor that wanted a screening technology to identify stock that had started to go bad. Other clients, concerned about receiving the wrong chemicals from suppliers, sought a way to check arriving shipments to prevent the contamination of storage tanks. The Cyranose, after being programmed for a specific application, is simple enough to be used by guard shack personnel.

The medical potential is also promising. For centuries physicians and other health workers have claimed to know certain diseases by smell. Typhoid victims are said to exude the scent of freshly baked bread; a whiff of ammonia may indicate kidney failure. My father came to recognize the odor of measles during his years as a general practitioner. The e-nose is already confirming some of these beliefs. In a clinical trial under way at the University of Pennsylvania, physicians hope the Cyranose can serve as an early-warning indicator of pneumonia in patients on respirators. “Prelimmary data shows a high degree of correlation,” says Sunshine.

Since September 11 Cyrano has received dozens of calls from government agencies, some of which have purchased units. Exactly how the technology is being employed is impossible to know: The contact the company normally has with commercial clients is nonexistent with the authorities. They come in, they buy, they disappear. The Cyranose does detect the nerve gas sarin, which killed 12 people in the Tokyo subway in 1995, and is even more adept at sniffing out Vx gas, the most lethal airborne chemical agent.

Biological weapons are still invisible to any electronic nose. Like all living organisms, bacteria such as Bacillus anthracis, which causes anthrax, emit telltale hydrocarbons, but the cells must be cultured first to create a measurable quantity. Direct detection–such as waving an instrument over a suspicious white powder and getting an instant warning beep or all-clear chime–is the Holy Grail of chemical-sensor technology, says Sunshine.

In one of several large, workshop-like testing rooms, Sunshine pauses at a table where a unit is being schooled to recognize a collection of household scents, from common fragrances to rubbing alcohol. He jams a small vial onto the unit’s needle probe and presses the yellow RUN button. In less than 30 seconds, the word OBSESSION appears on the LCD display.

The marketing limitations become clear when the same unit is prompted to sample the open air in the lab. As Sunshine predicts, the words UNKNOWN OUTLIER appear. The mostly colorless and odorless gas we breathe, of course, is full of compounds and other substances–especially in smoggy Pasadena–in both massive and minute quantities. To the Cyranose, however, it’s floating static. The sensors ignore light substances like nitrogen, oxygen, and carbon dioxide and aren’t designed to itemize the rest of the soup. Herein lies the somewhat superficial nature of the e-nose: It takes a 32-dimensional snapshot of any given mixture without divining the ingredients–even though it could recognize some of them if they were isolated. Like most connoisseurs, the Cyranose easily pegged Calvin Klein’s Obsession with a single sniff but doesn’t know what makes up the stuff. In scientific terms, it doesn’t speciate.

“If you can’t speciate, you can’t calibrate,” says Ed Staples, managing director of Electronic Sensor Technology, a Ventura County firm that makes an electronic nose using a different technology. Staples named his machine the zNose, and one soon gets the feeling he deliberately picked the letter farthest from e to show his disdain for the competition and its origins. “Professors,” he snorts. “Caltech is part of the problem. Nate Lewis will sell you a bill of goods.”

The zNose samples odors using gas chromatography, a 50-year-old technology for splitting a substance into its constituent parts. Whereas a traditional gas chromatograph takes several minutes to analyze a sample and is the size of a small refrigerator, the zNose works in seconds and comes in both bench-top and portable models. The latter vaguely resembles a Dust-buster connected to a laptop computer and is still much bulkier than the Cyranose, weighing 35 pounds.

While the Cyranose’s sensor array reacts to a substance all at once, the zNose’s single detector captures and weighs the individual parts, compound by compound, from lightest to heaviest. One by one the substance’s component hydrocarbons land on a crystal detector, where they briefly change its frequency before returning to the surrounding air.

The difference between the e-nose and the zNose is neatly personified by the smooth, youthful Sunshine and the macho, bristle-mustached Staples, who looks a good 20 years his counterpart’s senior. While the Cyranose can sort out Calvin Klein’s most popular perfumes, the zNose has analyzed new, car smell and sniffed the barrel of a recently fired .44 Magnum. During a two-hour meeting Sunshine never once resorts to aphorism to make his case; Staples repeatedly says the zNose gives scientists “one eye in the kingdom of the blind.”

Unlike the Cyranose, the zNose measures the concentration of ingredients. It doesn’t go as far as naming each constituent, however. The resulting data is about as meaningful to a novice observer as a Richter trace after a temblor, but a trained eye would quickly spot the presence of something that didn’t belong or the absence of something that did. A materials specialist at an automaker could inspect a batch of interior plastic for impurities, thus sparing the assembly line a potential run of foul-smelling lemons. An inspector at an orange juice-processing facility could use the zNose to sniff storage tanks for the buttery smell of diacetyl, a sign that the contents have gone bad.

The field of food quality and safety, in fact, is where the zNose thrives. Winemaker Sutter Home uses it to check corks for trichloral anisol, which can spoil a bottle’s contents in concentrations as low as a few parts per trillion. Johnnie Walker uses the zNose to sniff out counterfeit whiskey. “An electronic nose must be fast, accurate, and precise,” Staples says, uncapping a bottle of Mau Tai liquor, whose bouquet is vaguely fruity at first and then turns sour after about three seconds. “See? Humans are actually gas chromatographs, too.” By this he means that our noses, like the zNose, deal with complex smells sequentially, as heavy molecules take longer to reach the olfactory epithelia than lighter ones.

Despite the parallel, he insists that the zNose is purely an analytical instrument–and that the e-nose isn’t. “That’s for biology guys, A.I. guys. We don’t worry about neural nets and this n-dimensional space stuff. We can smell drinking water and tell you if it meets EPA standards.” The zNose is the only electronic nose validated by the Environmental Protection Agency, and the City of Diamond Bar has considered putting a unit in a van to help catch air polluters. “With smell there’s no boogeyman,” Staples says. “It’s straight science.”

Staples hedges when asked about its usefulness against chemical weapons. It would entail too many false alarms, he says. Even burning oil has the same basic hydrocarbon structure as nerve gas. “A lot of these government guys aren’t thinking straight,” he says. “We’re not a bomb detector.” Technical limitations aren’t the only problem: In 1996 he sent two units to the FAA, where the boxes sat unopened for three years. One airline, meanwhile, wants to use the zNose to analyze cockpit fumes, and Disney-world officials once considered having a phalanx of zNoses sniff incoming park visitors.

Law enforcement seems more interested in finding money, says Staples. At this, the zNose is a peerless bloodhound. During a single morning of probing sealed packages at the LAX post office, a unit located more than $250,000 by sniffing out the vapor from the money’s ink. When asked if the zNose could tell real money from counterfeit, Staples balks at the question’s phrasing–the artificial intelligence implications again. It doesn’t make decisions. “Sure, if the chemistry is different,” he explains. “It’s all chemistry.”

Staples plans to market a consumer version of the zNose, which currently fetches about $20,000, within the next year and sell it for less than $1,000. How unskilled owners would operate a handheld gas chromatograph is uncertain, as is the matter of what they would do with it. A diaper sniffer for the baby? A kitchen alarm for the cookies in the oven? A garden monitor to screen for the vapors plants release when attacked by insects? “We want it to be a nose, but a good nose,” he says, adding that other e-nose makers have “muddied the water.”

Cyranose’s Sunshine isn’t surprised by Staples’s opinion. “We don’t purport to have an analytical instrument,” he says. “It’s a different tool. I think there’s a place for both.”

Neither is ready for mass marketing. A version of the electronic nose might be integrated into a new model of refrigerator or oven before stand-alone units show up at Brookstone. “People think, Oh, you’re going to solve all my nose problems,” says Sunshine. “Eventually, `electronic nose’ will be the right name.”

ELECTRONIC-NOSE TECHNOLOGY is not bionic, nor is it meant to be. For the small number of Americans born every year with anosmia–the total absence of smell–there is no nasal equivalent of a hearing aid or LASIK surgery. Those suffering from the disease, which runs in families, can’t hook themselves up to an e-nose; they often face both their parents’ shame and their friends’ incredulity.

Since smell accounts for as much as 90 percent of what we perceive as taste, an anosmic experiences the world without two of the body’s five routes of input and is frequently oblivious to impending danger. “As a child I fell asleep while cooking supper and was awakened by the neighbor–not by the black smoke above my head,” wrote one victim on an Internet bulletin board last year. “Taxidermy gone awry has decomposed in my room, I’ve had food poisoning numerous times, and I’ve been to the hospital more than once because of ammonia poisoning. Having no sense of smell is dangerous in this Western society of ours.”

Others may experience anosmia or related disorders at some point during life. Severe head trauma can sever the connection between the olfactory epithelia and bulb. A rare form of brain tumor can press directly on the bulb, causing olfactory hallucinations. Workers in nail salons and other chemically harsh environments sometimes experience hyposmia, a dulling of smell, as do smokers and cocaine users. German physiologist Ernst Weber damaged the receptors in his nose when he bent over and poured perfume into his nostrils in an effort to prove that only gases, not liquids, can be smelled. (He was wrong.)

Like the other senses, smell deteriorates with age. It shouldn’t disappear altogether, however, and Jack Hilton has come to the Nasal Dysfunction Clinic at UCSD’s Thornton Hospital in La Jolla from Green Valley, Arizona, because one day last January he couldn’t smell the popcorn he was popping. Since then Hilton, 84, hasn’t experienced a single odor. He lives in a retirement home and isn’t concerned about missing a fire or a gas leak.

He does miss the taste of foods like chocolate. “You get to depending more on texture,” he says; he now finds himself eating lots of prunes and saltine crackers. A former FAA administrator, Hilton is articulate and healthy, looking, his round head wrinkled like soft leather and his mouth encircled by a white goatee. He fears lapsing into the sensory-deprived torpor in which many of his fellow residents exist. “Some of them just live in cocoons,” he says.

Hilton is questioned about his medical history by William Cain, a professor of otolaryngology and the director of UCSD’s Chemosensory Perception Laboratory. A clear pattern soon emerges. Around 1930 Hilton’s sinuses became so congested that a physician was forced to puncture and drain them. Since then he has endured occasional infections and postnasal drip. He has moderate to severe allergies. About five years ago he had a polyp removed, receiving postoperative care that Cain finds lacking.

Cain hands Hilton small containers of household products such as cinnamon, peanut butter, and Ivory soap and shows him a list of the smells to help him identify them. Hilton gets just a hint of the mothballs and Vick’s Vapo-Rub but otherwise can’t discern anything. He is then given a series of plastic bottles with the alcohol butanol in increasing concentrations. A healthy nose would pick up the liquids vapor even at the weakest dilution, but not until one of the last bottles–the nasal equivalent of a baseball bat to the head–does Hilton notice a tingling sensation.

Not being a medical doctor, Cain relies on the clinic’s physician, Terence Davidson, to make a diagnosis but suspects Hilton suffers from sinus disease, a frequent and often undetected cause of anosmia. The treatment in such cases is usually pharmaceutical: steroids, saline rinse, more steroids. Hilton stands a good chance of smelling popcorn again.

Leaving the clinic for the day, Cain voices disapproval of modern medicine’s handling of anosmia victims. “I get the feeling the physicians aren’t listening,” he says. Informative histories, like Jack Hilton’s lifetime of sinus trouble, are being ignored in favor of clinical testing, which may be inconclusive. It takes more than ten minutes to get information from a patient,” Cain says. He even admits to periodic disagreements with Davidson, especially in cases where examinations such as CT scan and endoscopy don’t support the pattern and history that Cain uncovers. “His evidence trumps mine.”

If some doctors can’t grasp the plight of anosmics, Proust could. In 1891, at the age of 20, he was prompted to fill out a personal questionnaire at a social event. When asked to describe his dream of happiness, he demurred carefully. “I don’t dare say,” he wrote. “I fear destroying it with the mere mention.” The perfect scent can be just as fragile and slippery; one fears inhaling too deeply will spoil the effect. In a fleeting wisp of air a tiny fantasy world–or a lost love, or an entire childhood–may rush back into the mind’s arms or vanish forever into the night’s ether. For someone who has lost the sense of small, these are times never regained.

Senior editor Joshua Tompkins (“The Vapor Chase,” page 48) writes about science and technology for Los Angeles. His story about Argentine ants appeared in the February issue.

Juliette Borda’s paintings (“California or Bust,” “The Skin You’re In,” “The Vapor Chase,” pages 38-51) have appeared in Esquire, The Atlantic Monthly, and The New York Times. Borda’s work has been exhibited at the Norman Rockwell Museum in Stockbridge, Massachusetts, and is in the permanent collection of the Carnegie Museum of Art in Pittsburgh.

Dave Lauridsen’s photographs (“Lord of the Rink,” page 28) have appeared in the Los Angeles Times Magazine. He photographed R.M. Schindler’s classic Tischler House for the February issue.

COPYRIGHT 2002 Los Angeles Magazine, Inc.

COPYRIGHT 2002 Gale Group