If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Analytical Chemistry

Food Detectives

Researchers apply advanced analytical techniques to profile naturally occurring chemicals to track food origin, nutrition, and flavor

by Jyllian Kemsley
October 8, 2012 | A version of this story appeared in Volume 90, Issue 41

Credit: Martin Steinhaus
Technical University of Munich graduate student Julia Wagner detects scent compounds in roasted hazelnut extract using GC olfactometry.
Technical University of Munich graduate student Julia Wagner detects scent compounds in GC olfactometry of a roasted hazelnut extract.
Credit: Martin Steinhaus
Technical University of Munich graduate student Julia Wagner detects scent compounds in roasted hazelnut extract using GC olfactometry.

From spinach to strawberries, honey to tomato sauce, food is a complex mixture of chemicals. Whether you’re looking at a single piece of fruit or a can of beef stew, each may contain thousands of compounds.

Food analysts historically looked for only a handful of these chemicals, such as residual pesticides, now advances in analytical techniques allow scientists to profile the entire chemical composition of a food. Used as something of a fingerprint, chemical profiles—not all compounds need to be individually identified—can be used to track the geographic origin of a vegetable or how a food was processed. Food manufacturers are using the information to ensure the safety and quality of their supply lines, especially for foods that command a premium price on the basis of factors such as where they were made or how long they were aged.

Further analysis of food chemical profiles can also reveal key markers for nutrient content, flavor, and aroma. The data allow for better understanding of the characteristics of different cultivars of fruits or vegetables and how they may best be stored or processed. In the past, the focus of plant breeders was often to improve crop yields through pest, disease, or drought resistance or to ease storage, transport, or processing, says Susan E. Ebeler, a professor of viticulture and enology at the University of California, Davis. Now, “what we’re really trying to do is bring flavor back to the forefront,” Ebeler says. “All those other things are important—we have to be able to store and transport. But if we want people to eat more fruits and vegetables, they also have to taste good.”

The key chemicals in foods that determine their unique fingerprints or give qualities such as flavor tend to be the compounds known as secondary metabolites. Primary metabolites are compounds that are necessary for plants’ growth, development, or reproduction. Secondary metabolites are compounds that may not be absolutely required for plant life but still play a role in plants’ survival. Examples include the pigments or aromas plants use to attract pollinators or the toxins they wield against predators.

To profile all of the compounds in a food, researchers turn to analytical techniques such as gas or liquid chromatography combined with mass spectrometry (GC/MS or LC/MS) or nuclear magnetic resonance (NMR) spectroscopy.

These analyses generate a lot of complex data, says Colin Thurston, director of informatics product strategy for Thermo Fisher Scientific. As a result, he says, scientists must use informatics approaches to pick out key differences in the data that would lead them to narrow in on one or a few chemicals to track. Once that subset is identified, manufacturers can figure out what techniques are best suited to monitor them in the field or a quality-control lab. Thermo Fisher has one customer that uses an infrared scan to measure the maturity of aged beef and determine whether it can be released to market, Thurston says.

Roberto Consonni, a researcher at Italy’s Institute for Macromolecular Studies, is one scientist actively pursuing food metabolite profiling, with NMR as his tool of choice. Although Consonni’s published work focuses on food authentication for fraud detection, he also does process analysis for nutrients in private projects for companies.

On the authentication side, Consonni has used NMR to probe the geographic origin of honeys. According to an analysis by Global Industry Analysts, the global market for honey will exceed 1.9 million tons by 2015. Honey is also the third-most-adulterated food, after olive oil and milk, as determined by a study of the U.S. Pharmacopeial Convention’s Food Fraud Database, which includes fraud reports from English-language publications from 1980 to 2010 (J. Food Sci., DOI: 10.1111/j.1750-3841.2012.02657.x). Unscrupulous honey makers have been known to dope their product with other sweeteners, such as high-fructose corn syrup, or to substitute a less expensive honey from a different geographical or botanical origin for a high-end honey.

Consonni and coworkers used patterns of NMR signals from sugars to distinguish honey samples from China, South America, Hungary, and Italy. Consonni’s group has also used sugar NMR signals to further differentiate Italian honeys. The sugar compositions differ depending on whether the honeys were made from plants grown at low or high altitude and whether the pollen came from an assortment of flowers or more specifically from rhododendron plants, acacia trees, or chestnut trees (J. Agric. Food Chem., DOI: 10.1021/jf3008713).

Consonni has also used NMR to track how expensive cheese was processed or aged. As cheese ripens, for example, its proteins, lipids, and sugars undergo chemical and physical changes. Under European Union regulations, Parmigiano-Reggiano cheese can be given that name only if it is produced in a certain region of Italy by a particular process. The worldwide market for Parmigiano-Reggiano is about $1.3 billion, Consonni says. Because of the high price of this cheese, especially if it’s aged for two to three years, manufacturers or retailers may try to substitute less expensive cheese or label young Parmigiano-Reggiano as older.

In a 2008 study, Consonni and coworkers found that NMR could distinguish Parmigiano-Reggiano cheese aged for 14, 24, or 30 months (Talanta, DOI: 10.1016/j.talanta.2008.02.022). Homing in on the specific chemicals and their variations, the researchers found that younger samples contained more leucine and isoleucine, whereas older samples contained more threonine. They also used the technique to distinguish Parmigiano-Reggiano from similar hard cheeses from Eastern Europe. The Eastern European cheeses contained more leucine, isoleucine, lactic acid, butanoate, and acetic acid relative to Parmigiano-Reggiano, whereas the Parmigiano-Reggiano samples had comparatively higher amounts of threonine, valine, proline, glutamic acid, lysine, alanine, serine, arginine, and citrulline.

Wine compounds may also vary and change depending on the grapes from which they’re made and how they are fermented and stored. UC Davis’ Ebeler uses MS techniques to study wine. A new food safety and measurement laboratory, established at UC Davis with support from instrument firms Agilent and Gerstel and wine, beer, and spirits purveyor Constellation Brands, will help that effort. Ebeler and lab codirector Alyson E. Mitchell, a professor of food science and technology at UC Davis, hope that the new facility will serve as a research hub on campus for studies of food authentication and will increase understanding of plant secondary metabolism, microbial metabolism, and food flavor.

In one project, Ebeler and colleagues are using LC/MS to fingerprint different wines and identify the grapes that were used to make them. Although the project is still in the early stages, they’ve built enough of a spectral library to be able to separate and identify wines by their components. As samples come in, sometimes there’s an outlier—and when the researchers ask the winery to check its records, the winery generally turns up a labeling error. “Winemakers don’t believe that the analytical tools are that good, but they are,” Ebeler says.

Along with colleague Hildegarde Heymann, a professor of sensory science in the department of viticulture and enology at UC Davis, Ebeler has also looked at the effects of shipping conditions on wines by storing bottles at a constant temperature of 20 or 40 ºC, by cycling bottles between the two temperatures to reflect night/day cycles, and by storing some bottles in the trunk of a car. Storage at higher temperature had the biggest effects on both red and white wines. White wines generally gained 1,1,6-tri­methyl-1,2-dihydro­naphtha­lene and vitispirane 1 and 2 while losing isoamyl acetate, hexyl acetate, and 2-phenylethyl acetate; wine tasters documented that the wines picked up “diesel, oxidized, and rubber” aromas and moved away from “citrus, floral, and tropical fruit” (Am. J. Enol. Vitic.2010, 61, 337). The red wines also showed chemical changes, with decreased amounts of a variety of compounds, but the wine-tasting panel did not note significant differences.

Ebeler’s colleague Mitchell has also looked at the effects of storage conditions on food, in particular flavonoids in onion. Flavonoid-rich foods may help prevent cardiovascular disease and cancer. Mitchell and postdoctoral researcher Jihyun Lee found that the Milestone variety of onion contains the most flavonoids of 10 cultivars studied (J. Agric. Food Chem., DOI: 10.1021/jf1033587). Tracking one particular flavonoid, quercetin 4′-O-glucoside, in freeze-dried and powdered onion, Mitchell and Lee found that the amount of quercetin decreased in the first couple of months of storage, then stabilized for the following 10 months. They also found that significant amounts of flavonoids are discarded in onion-processing waste products, opening up the possibility of using those products as a source of flavonoids for dietary supplements and cosmetics.

Mitchell and colleagues also looked at flavonoids, along with ascorbic acid and nitrate, in 27 varieties of spinach grown under organic and conventional farming conditions. Although flavonoids and ascorbic acid (vitamin C) are good for health, nitrate can be detrimental. The researchers identified 17 flavonoids in the spinach samples and found that the amounts of flavonoids and ascorbic acid were higher in organically grown spinach (J. Agric. Food Chem., DOI: 10.1021/jf300051f). Conventionally grown spinach, in contrast, was higher in nitrate. The increase in nitrate amounts was likely because of its availability through fertilizer use on conventionally farmed plots.

Other researchers are profiling plant metabolites as a starting point to pin down key aroma or flavor compounds in foods. The taste or odor of a food usually depends on an ensemble of compounds, says Peter Schieberle, director of the German Research Center for Food Chemistry and a professor of food chemistry at Technical University of Munich. Individual scents tied to individual compounds picked up by a human nose may smell nothing like the food being analyzed. “Let’s say you have 25 compounds, and none of them smells like roast beef,” Schieberle posits. “But when you mix them together in the right quantitative amounts, immediately the specific aroma of ‘roast beef’ comes up.” Consumers are demanding foods they like year-round—not just when particular fruits or vegetables are in season—and wanting processed foods to mimic those prepared in individuals’ kitchens. So food producers increasingly want to know what the ideal “flavor blueprint” of a food is so they can create or maintain it, Schieberle says.

A common way to analyze for odor compounds is to run a sample through a GC column, then split the effluent. One part of the effluent goes to an instrumental detector, such as a flame ionization detector or mass spectrometer. The other goes to a person—or, more specifically, their nose. The human detector notes what they smell at particular times, and that scent gets correlated to whatever the instrument detector picks up.

In one study, Schieberle and colleagues looked for key odor-active compounds in raw Italian hazelnuts and roasted-hazelnut paste. Schieberle’s group identified 19 odorants in raw hazelnuts, 15 of which were present in concentrations high enough to be smelled. Of those 15, flowery linalool, nutty and fruity 5-methyl-4-heptanone, and earthy 2-methoxy-3,5-di­methyl­pyrazine are the strongest drivers of raw hazelnut scent (J. Agric. Food Chem., DOI: 10.1021/jf300908d). For roasted-hazelnut paste, the group identified 25 odorants, of which malty 3-methyl­butanal, buttery 2,3-pentane­dione, popcornlike 2-acetyl-1-pyrroline, tallowy (Z)-2-nonenal, and cabbage-smelling dimethyl tri­sulfide mainly create the odor of the paste. Degradation of free amino acids present in the raw nuts yields the roasted-scent compounds.



Schieberle and postdoctoral researcher Monika Christlbauer have also looked for aroma compounds in beef and pork vegetable gravies. They prepared the gravies by browning meat cubes in lard, removing the meat and browning the vegetables, and then combining everything and stewing it in an oven at low heat for several hours before removing the meat and blending the remaining liquid. Their analysis, published in 2011, showed that the key aroma compounds were oniony 3-mercapto-2-methyl­pentan-1-ol, fatty (E,E)-2,4-deca­dienal, cucumberlike (E,Z)-2,6-nona­dienal, fatty- and grassy-smelling (E)-2-decenal, and soapy (E)-2-undecenal. In beef gravy, 12-methyl­tri­decanal provided a tallowy note, whereas (E,Z)-2,4-deca­dienal gave more of a fatty aroma to pork gravy (J. Agric. Food. Chem., DOI: 10.1021/jf203340a). Food manufacturers can use the information to help create ready-made gravies that more closely resemble homemade, Schieberle says.

Russell L. Rouseff, a professor of food science at the University of Florida Citrus Research & Education Center, has recently been studying strawberries and blue­berries to try to identify important odor compounds and how well they’re expressed in different cultivars of berry plants, especially because odor is an important component of flavor. Fruits tend to produce flavor compounds late in life, so part of the effort involves finding the “sweet spot” to harvest berries when they are just ripe enough to develop their full flavor but not so ripe that shelf life becomes a problem.

One of his projects involves understanding the aromas and flavors of two strawberry cultivars developed by the University of Florida to grow in Florida during winter months, when California production drops off. Florida Radiance is adapted to grow in November, December, and March, whereas Strawberry Festival grows best in January and February. Russell’s 2011 analysis of volatile aroma components of the two strawberry cultivars using standard GC techniques showed that although they contain some aroma compounds observed in other cultivars, they have a different aroma pattern from summer-grown strawberries. They also contain three aroma compounds that other cultivars do not: grassy- and cucumber-scented (E,Z)-2,6-nona­dienal, sweet- and berry-scented 3,7-di­methyl­octa-2,6-dien-1-ol, and fruit- and pineapple-scented ethyl hexanoate (J. Agric. Food. Chem., DOI: 10.1021/jf2030924).

Standard GC techniques can miss sulfur compounds, which Rouseff calls “the stealth volatiles of the flavor world.” Sulfur compounds likely arise as reaction or degradation products of sulfur-containing amino acids. The human nose can smell very small amounts that many analytical techniques, including MS, cannot detect. To find them, Rouseff combines GC with a special pulsed-flame photometric detector with filters tuned to pick up sulfur emissions.

Using that technique to study Strawberry Festival and Florida Radiance, he and colleagues in 2011 identified methyl thioacetate, methyl thiobutyrate, and ethyl (methylthio)acetate as the top three aroma-producing sulfur compounds in the cultivars (J. Agric. Food Chem., DOI: 10.1021/jf104287b). Ethyl (methylthio)acetate had not previously been identified as an aroma compound. The researchers also found that those and other sulfur compounds were generally not present in immature fruit and increased in concentration as berries ripened. With individual scents akin to cheese, garlic, cabbage, and onion, the compounds may not seem to smell much like strawberries, but they nonetheless may contribute to the overall aroma of ripe fruit, Rouseff says.

Food metabolite profiling, for whatever purpose, is a field in its infancy, says Steve Royce, Agilent’s food segment marketing manager for the Americas. It is important economically and it has “islands of very highly evolved science,” Royce says, but big knowledge gaps still exist. For example, if you want to pinpoint the kind of meat that is in a can of soup, right now “there’s no analytical technique that can give you that kind of information,” Royce says. The field is so broad, it is “nirvana” for instrument makers, he says, because everything an instrument firm makes is going to be used as the food industry tracks its products more closely.

Clearly, it is an area primed for further exploration: for the development of new methods and perhaps instrumentation to make our food supply more transparent, healthier, and more flavorful.


This article has been sent to the following recipient:

Chemistry matters. Join us to get the news you need.