Adding milk to coffee changes not only its bitterness and appearance but also its aroma during drinking. In a 2015 study, coffee-like aroma was perceived as weaker with milk, yet many aroma compounds in the air exhaled through the nose showed no significant difference in amount. Changes were found in only some compounds.
Using the compounds that decreased as a clue, the researchers supplemented two aroma compounds and confirmed an increase in coffee-like aroma. Rather than treating the whole aroma as uniformly diluted, the study explored the possibility that milk changes the release of particular important compounds.
Measuring breath after drinking, rather than aroma above the cup
Aroma is perceived not only by sniffing through the nose, but also through the route from the back of the mouth and throat to the nose after food or drink enters the mouth. This study collected air exhaled through the nose after coffee consumption to examine which compounds emerged and in what amounts.
The researchers compared black coffee made from 100 g of coffee extract plus 100 g of water with milk coffee made from 100 g of extract plus 100 g of cow’s milk. The extract was prepared from medium-roasted Colombian Arabica beans. The amount of coffee extract was matched, so the milk condition was not simply more diluted. The commercial milk contained at least 3.5% milk fat, and the beverages were served at room temperature.
Three trained participants took 30 mL into their mouths, swallowed after 1–2 seconds, and had their next ten breaths collected. Each condition was measured seven times per person, and the compounds were compared using gas chromatography–mass spectrometry. These breath measurements are not the instantaneous concentrations reaching olfactory receptors in the nose, but they provide information about aroma released from the mouth during consumption.
Most of 39 compounds showed no significant difference; three decreased in all three people
The amounts released of the 39 analyzable compounds varied between individuals. For many compounds, the difference between black and milk coffee was not significant. Three compounds, however, decreased significantly in all three participants: furfuryl methyl sulfide (FMS), difurylmethane, and furfurylpyrrole.
For FMS in particular, the amounts with milk were 0.52, 0.48, and 0.51 for the three participants when the amount with black coffee was set to 1: roughly half as much. Other compounds decreased in some participants, indicating that milk’s effects differed by both compound and person.
The three compounds that decreased were relatively hydrophobic, consistent with previous research showing that milk fat affects aroma release. Some other similarly hydrophobic compounds, however, did not decrease significantly. Affinity for fat alone cannot explain all the findings.
Using measured FMS to consider the behavior of important FFT
The researchers also focused on 2-furfurylthiol (FFT), an important contributor to coffee aroma. Previous research had reported that FFT is converted to FMS during passage through the mouth. The FMS exhaled through the nose may therefore include both FMS originally present in the beverage and FMS formed from FFT.
The reduction in FMS with milk in this study provides a clue that FFT release might also have decreased. It is not, however, a direct measurement of the amount by which FFT decreased.
Alongside the effects of milk fat, the authors discussed the possibility that FFT reacts with milk proteins such as casein, making it less available for aroma release. This reaction pathway was not directly tested in the study; it is a hypothesis proposed to explain the measurements.
Supplementing two compounds enhanced coffee-like aroma
In the next stage, experiments including the addition of FMS to deodorized coffee were used to examine the release of pre-existing FMS and the portion thought to derive from FFT. Breath analyses for adjusting the addition levels were conducted with one trained participant.
Adding 80 ppb FFT and 10 ppb FMS to milk coffee produced a condition in which exhaled FMS approached the amount measured with black coffee. Next, 31 participants with sensory-training experience drank and evaluated this adjusted milk coffee, unadjusted milk coffee, and black coffee.
To reduce the influence of color, samples were presented under red lighting in a dark room. Coffee-like aroma intensity was compared on a seven-point scale, with the unadjusted milk coffee assigned a reference value of 4. Black coffee was stronger than unadjusted milk coffee, and milk coffee supplemented with the compounds was also significantly stronger than the unadjusted version. No significant difference was detected between black coffee and the adjusted version.
The result supports the idea that supplementing the delivery of certain compounds can change perceived coffee-like aroma. Because FFT and FMS were added together, however, the comparison does not isolate the effect of FFT alone. It also assessed the intensity of coffee-like aroma, not preference or overall deliciousness.
Summary
For milk coffee, many of the 39 compounds exhaled through the nose showed no significant difference, but three, including FMS, were lower than with black coffee in all three participants. Milk coffee supplemented with FFT and FMS was also perceived as having a stronger coffee-like aroma than the unadjusted version.
A beverage’s aroma depends not only on the list of compounds in the liquid but also on changes in the mouth and release along the route to the nose. These findings offer a way to consider milk-related aroma changes through the behavior of a few important compounds. Further work is needed to establish what happens with hot coffee, other milks, and more people.
The three authors and the 31 sensory assessors were affiliated with Ogawa & Co., Ltd. at the time of the research. No separate funding or conflict-of-interest declaration section was identified.
References
Itobe, T., Nishimura, O., & Kumazawa, K. (2015). Influence of Milk on Aroma Release and Aroma Perception during Consumption of Coffee Beverages. Food Science and Technology Research, 21(4), 607–614. doi:10.3136/fstr.21.607 (Full original article)
