How Does Nut Milk Change Coffee Flavor?

Replacing the milk added to coffee with a nut-based alternative changes not only aroma but also perceived sweetness and mouthfeel. Do these differences also depend on which nut is used and how it is processed?

In 2025, Woorin Lee and colleagues prepared milk-like beverages from four types of nut, mixed them with the same espresso, and compared them. Their aim was to investigate how nut type and roasting affected composition, physical properties, and sensory ratings, providing clues for improving plant-based coffee drinks.

Four nuts, compared with and without roasting

The ingredients were almonds, cashews, hazelnuts, and walnuts. Each was prepared unroasted or roasted at 140°C for 15 minutes. Nuts were blended with water at a 1:9 weight ratio and filtered through cloth to make the milks. This was not a comparison of commercially available plant-based milks.

The researchers added 150 mL of nut milk to 50 mL of espresso made from Brazilian Arabica. They also prepared coffee with commercial dairy milk for comparison, and some compositional analyses used espresso diluted with water at the same ratio as a reference.

Measurements included volatile compounds, fatty acids, color, particle size, viscosity, and in-vitro antioxidant activity. Sensory evaluation was performed by 17 trained panelists aged 22–28 at Dongguk University in South Korea. Samples were identified by codes and rated for attributes including overall liking and mouthfeel.

Dairy milk had the highest mean liking, with cashew relatively high among nuts

On an overall scale from 1, extremely dislike, to 9, extremely like, dairy-milk coffee had the highest mean score, 6.12. Among nut-based samples, unroasted cashew scored 5.24. Mean scores for the other nut samples ranged from 3.24 to 4.53.

However, the order of means does not automatically establish reliable superiority. The statistical notation in the original table does not indicate a significant difference between dairy milk and unroasted cashew. Nor were there significant differences between unroasted cashew and many other nut samples. The findings identify cashew as a promising candidate, not a proven winner over every alternative.

Dairy-milk coffee also had the highest mean mouthfeel-liking score. Improving nut-based coffee may therefore require more than adjusting aroma; smoothness and other beverage properties also warrant attention.

The original paper describes ratings for individual tastes as liking scales in Methods but interprets some as taste intensity in Results. This limits interpretation of the strength of bitterness and nuttiness.

Nut aromas overlap with coffee aromas

The volatile analysis reported 33 compounds. Nut-based coffees had higher measured values for compounds associated with nut-like aromas, including benzaldehyde and some pyrroles. Dairy-milk coffee had higher values for compounds including 5-methylfurfural and furfuryl acetate. The ingredients mixed into coffee thus changed its volatile profile.

These comparisons used analytical peak-area ratios relative to an internal standard. They did not directly measure how strongly a particular compound was perceived while drinking. Volatile release changed alongside the ingredient composition and the overall beverage matrix, including fats and proteins.

Roasting effects were not simple either. For example, roasted-almond samples contained less benzaldehyde than unroasted-almond samples. Roasted nuts also produced darker beverages with higher browning indices and lower viscosity. Roasting was not merely a way to add toasted aroma: it changed both composition and beverage properties.

Higher compositional values do not necessarily mean greater liking

Fatty-acid analysis reported higher amounts of unsaturated fatty acids in nut-based samples than in dairy-based coffee. In the DPPH test, which measures radical-scavenging capacity in vitro, roasted-nut samples had higher values than their corresponding unroasted samples. Walnut-based samples had high total phenolic content, with differences among ingredients.

Based on these properties, the authors discuss potential for plant-based beverages with nutritional considerations. However, the study did not follow the health of people drinking them. In-vitro antioxidant activity and fatty-acid measurements alone do not establish health benefits or superiority to dairy milk for health.

Compositional characteristics and liking were also distinct outcomes. Even when roasting increased measured antioxidant activity, it did not uniformly improve overall liking. Product development needs to consider both desired compositional properties and flavor and mouthfeel.

The authors propose improving flavor and mouthfeel together

When the researchers examined results across samples, aldehyde measurements were negatively correlated with overall liking, while mouthfeel liking was strongly positively correlated with overall ratings. On this basis, the authors propose controlling formation of compounds associated with less-liked bitterness and aftertaste while also improving particles and beverage structure.

Correlation does not establish causation, however. Whether reducing a particular aldehyde alone increases liking, or reducing particle size produces similar improvements, must be tested by actually changing formulations. This study identifies compositional and sensory features to focus on in subsequent prototypes.

The findings concern 17 people, specific formulations, and limited origins for each ingredient. Commercial products may use other approaches, including sweetening, emulsification, and homogenization, so the study’s ranking cannot simply become a shopping ranking. Plant-based drinks not tested here, such as oat and soy, are also outside its scope.

The study received support including the National Research Foundation of Korea’s Basic Science Research Program, and the authors declared no conflicts of interest.

Summary

In nut-milk coffees, ingredient choice and roasting changed not only volatile compounds but also color, viscosity, and sensory ratings. Dairy milk had the highest mean overall rating in these samples, while unroasted cashew rated relatively highly among the nut-based alternatives.

The authors’ goal is to create beverages that make use of nut-derived compositional properties while offering acceptable flavor and mouthfeel. This comparison suggests that doing so involves more than switching ingredients: aroma-related compounds and beverage structure can be adjusted together.

References

Lee, W., Ahn, H., Yim, J., Kim, Y., & Lee, K.-G. (2025). Physicochemical properties and sensory attributes of nut-based milk coffee. Scientific Reports, 15, 24238. DOI: 10.1038/s41598-025-07554-w · Full original article.

Original article: © The Author(s) 2025. CC BY-NC-ND 4.0.