What the Nose Knows - Avery Gilbert [13]
At the heart of the GC is a Slinky-esque coil of very thin tubing that would stretch ten to thirty meters if unwound. As a first step, the smell sample is injected into the coil, where it is absorbed into a polymer that coats the inside of the tube. The Slinky sits in a little oven, which heats up in preprogrammed steps over the course of two minutes to two hours, depending on the setup. A stream of helium gas enters one end of the coil and exits the other. As the temperature rises, odor molecules are driven out of the polymer and into the gas stream. The process is orderly: each type of molecule evaporates and enters the stream at a specific temperature, depending on its molecular weight, and emerges from the end of the coil in a burst roughly two seconds long. The amount of material in each burst shows up as a peak on a timeline. The more molecules, the bigger the peak. A pure sample of a single chemical, say phenylethyl alcohol, yields a single peak. A complex mixture like rose oil produces a series of peaks, varying in height, representing the more and less plentiful components in the mixture.
Because it is highly detailed and unique to each sample, the visual profile created by the GC is often likened to a fingerprint. The difference is that a fingerprint is static—a direct physical impression—while the GC is dynamic: it takes a complex smell and pulls it apart in time. Perfumers liken a smell to a musical chord; if this is the case, then the GC plays it as an arpeggio.
As individual odors emerge from the GC, they can be fed into another device called a mass spectrometer, which provides a definitive identification of the molecule. By the mid-1970s the GC/MS linkage had been automated and labs around the world were churning out detailed chemical analyses of natural products. This was a mixed blessing for smell scientists. Run orange pulp through a GC/MS and you get a laundry list of volatile components. Do they all smell? Do they all contribute to the total orange aroma? How can we tell?
Since the early days of GC, chemists have sniffed at the exiting gas stream to see if they could recognize the emerging components by nose. Some volatiles, such as carbon monoxide, are entirely odorless to the human nose; otherwise each GC peak corresponds to a distinct smell. The size of the peak is not a reliable index of odor power. A big peak may deliver very little odor (which means the molecule is not very smelly) and a tiny peak may pack a punch (the molecule is a potent odorant). Cornell University chemist Terry Acree pioneered what is known as gas chromatography-olfactory or GC-O, which is essentially a formalized way of sniffing the GC vent to correlate smells with specific molecules. Acree devised a way to express numerically the relative odor potency of each chemical within a complex sample. He divides a chemical’s concentration in the sample by the minimum concentration needed to smell it on its own. Molecules with an odor impact index hovering around 1.0 are just at the level of detectability. Molecules with high multiples contribute more to the overall odor, while those with multiples less than 1.0 are seldom detectible; at best they lend a grace note to the overall composition.
Hey Beavis, Pull My Finger
One might expect the chemistry of certain bathroom malodors to be well understood. What other stinks are experienced on so personal a basis? For years, medical students were taught that the main ingredients of fecal odor were skatole and indole, nasty-smelling molecules created by the breakdown of meat protein during digestion. This claim persisted in textbooks despite never having been confirmed by direct chemical analysis. The shit finally hit the gas chromatograph in 1984 when researchers in Salt Lake City ran some poop through a GC and sniffed the results. Skatole and indole, although present in the sample, contributed relatively little to the typical fecal odor. The key actors turned out to be sulfur-containing compounds such as methyl mercaptan, dimethyl disulfide, and