Simulating the Coffee Drinker’s Nose
Is Scratch ‘n Sniff Starbucks in our future?
No industry focuses as much on olfactory marketing as the coffee business. Starbucks recently dumped its breakfast eggs because their smell didn’t pair well with the coffee aroma. Nestle unit Nespresso has not only modified its home brewing equipment to release more enticing smells, they have even launched a chain of coffee shops after finding that more than half of the coffee drinking experience came from the shop environment (see Sensory Marketing to Jolt Espresso Sales). Now, those clever coffee fanatics at Nestle have found way to analyze the components of coffee aromas that lets them predict how real human noses will respond to those smells.
Scientists at the Nestle Research Center in Lausanne, Switzerland, are reporting success in developing a system to judge the sensory qualities of a cup of espresso. Using a proton-transfer reaction mass spectrometer, which ionizes and analyzes the hot gases wafting above the coffee surface, the system can quickly predict what trained human tasters will say about it.
The aroma of roasted coffee contains as many as 1,000 volatile compounds, although a particular aroma can be defined and reproduced fairly accurately with about 50 or fewer. The system devised by Christian Lindinger and colleagues and described in the journal Analytical Chemistry does not rely on precisely identifying compounds but looks at how the mass spectrometry data differ from brew to brew. [From the New York Times – Scientists Finding Ways to Perfect a Cup of Joe, Without the Attitude by Henry Fountain]
The scientists tested the system by letting coffee tasting experts sample eleven different espressos and develop a sensory profile for each one. They also sampled the aromas with their mass spec equipment, and correlated the human analysis with the machine results. Then, they prepared a new batch of espressos, sampled the aromas mechanically, and predicted what human tasters would say. In fact, the predictions closely matched the actual human results.
Complex aromas like those of coffee and wine have been considered difficult or impossible to analyze and reproduce even with sophisticated equipment due to the large number of compounds they contain. In the paper in Analytical Chemistry describing this research, When Machine Tastes Coffee: Instrumental Approach To Predict the Sensory Profile of Espresso Coffee, one finds that a mere 16 key compounds were sufficient to be highly predictive of the human response. (For an entertaining take on this paper, complete with a poetic tribute to mechanized coffee analysis, check out The singularity can’t come soon enough at The Digital Cuttlefish.)
This work could have big implications for olfactory marketing – product quality and aroma variations could be tracked, environmental aromas could be characterized, and in some cases perhaps synthetic substitutes for even very complicated aromas could be developed.