Coffee Evaluation: Welcome to the Machine

Some food and beverage products rely heavily on human sensory evaluation. The coffee industry, in particular, practices the art of cupping, a tasting protocol, to evaluate the complex aroma and flavor profile of the brew for both quality control and development purposes. But much research has focused on how to reliably replicate results via machine, and researchers in Switzerland may have hit on the answer.

While not replacing the human touch and expertise needed for sensory evaluation, Christian Lindinger and colleagues at Nestlé Research, Vers-Chez-les-Blanc, have developed an instrumental approach that might augment the traditional methods, especially in quality-control situations for monitoring production and processing. Their findings, due to be published in Analytical Chemistry (When Machine Tastes Coffee: Instrumental Approach To Predict the Sensory Profile of Espresso Coffee, ASAP Article 10.1021/ac702196z S0003-2700(70)02196-0), found a robust and reproducible model to predict the sensory profile of espresso coffee from instrumental headspace data.

Despite the multisensory experience provided by drinking a cup of coffeecombining senses such as olfaction, taste, texture, trigeminal and visual sensationand the more than 1,000 compounds that may contribute to the experience of drinking coffee, the study found instrumental methods could assess the taste and aromatic qualities of espresso coffee nearly as accurately as a panel of trained human espresso tasters. Analyzing the gases released by a heated espresso sample, it transformed the most-relevant chemical information into taste descriptors like roasted, flowery, woody, toffee and acidity. The researchers combined GC/MS (electron impact, EI, ionization) with PTR-MS (chemical ionization) to build the bridge between molecular information and on-line marker based headspace data, applicable to complex food systems, says the report. After correlating the instrumental and sensory data from 11 espresso coffees, it was validated on a set of eight additional coffees.

The study authors concluded, as the model is based on fast on-line PTR-MS analysis, a prediction of a sensory profile can be accomplished within minutes. Relative to current methods of aroma profile analysis, this opens the possibility of high throughput studies.