Roast level and the chemical baseline

Roast level changes coffee's chemical baseline, so it can also affect the volatile-compound profile during storage. But “roast level” isn't a variable that can be separated from everything else. Storage temperature, time, whole-bean or ground form, and packaging can all change the rate of transformation. So to study the relationship between roast level and aroma loss, these variables need to be held within a controlled range first.

Compound profile during storage

A 2025 study on Thai Arabica found different roast levels have different volatile-compound profiles, and shelf life changes as storage conditions change. In that study, high storage temperature shortened shelf life and many volatile-compound groups decreased over time. This result shows roast level plays a role, but it isn't enough to create a simple rule like “darker roast means faster or slower aroma loss” applicable to every coffee.

Measurement and sensory perception

When tracking aroma loss, you need to distinguish compositional change from change humans perceive. A compound can decrease significantly analytically but have a small sensory effect if the concentration stays above or below a relevant threshold. Conversely, a small change in a strong-smelling compound can be noticeable. So volatile-compound analysis should be placed alongside sensory evaluation if the goal is a conclusion about drinker-perceived change.

Separating roast level from storage conditions

A comparison design can use the same coffee source, divide it into different roast levels, then pack and store at the same temperature. Sample over time to measure volatile-compound profile and a set of sensory indicators. When conditions are held stable, differences between the curves have a stronger basis for being linked to roast level. If both packaging and temperature change too, it becomes very hard to determine which part of the change belongs to roast level.

Limits when predicting shelf life

Storage temperature is especially important because it can change reaction and diffusion rates. Research on aroma loss in brewed coffee at different temperatures also shows the volatile-compound profile changes with storage conditions. So one roast level can behave differently at low versus high temperature. This also explains why you shouldn't take an accelerated test and directly convert the numbers to normal storage conditions.

The final limit is that shelf life can't be predicted from roast level alone. Roast level is an important explanatory variable, but post-roast quality also depends on packaging, temperature, oxygen, moisture, and product state. The safe way to interpret this is to say roast level can change the compound baseline and therefore interact with the aroma-loss process during storage; the specific rate must be measured under the actual conditions used.

A valuable design is keeping the same bean source and the same post-roast process, then creating multiple roast levels within a controlled range. Samples are packed identically and stored at multiple temperatures. This lets you check both the effect of roast level and its interaction with temperature. If only one temperature is used, it's hard to know whether the difference between roast levels is robust or just a feature of that specific condition.

You also need to distinguish whole-bean coffee from ground coffee. Grinding increases surface area and changes the rate of gas exchange, so the rate of aroma loss for grounds can't be assumed to match whole beans by default. If the shelf-life question focuses on roasted-ground product, the grind state must be held constant from the start. This is an example of how product scope determines how roast-level data should be interpreted.

The volatile-compound profile also shouldn't be read as an unfocused long list. You can choose groups or marker compounds already established as relevant to aroma, then track their change. If too many variables are included, it becomes easy to find a few random correlations. A design with a clear hypothesis helps limit interpretation after simply looking at the data.

When comparing roast levels, color can be a useful indicator for describing roast level but isn't itself a measure of aroma quality. Two samples with similar color can still differ in thermal history and compound profile. So if studying roast level, you should describe it using a suitable measurement method and keep roasting conditions controlled, rather than relying only on visual bean color.

Roast level should be described using a consistent measurement method rather than just labels like light, medium, or dark. These labels may not be fully equivalent across systems. When color or thermal-history data is available, they can help describe the sample group more clearly. However, even when roast level is well defined, the rate of aroma loss still depends on temperature, oxygen, packaging, and product state. So conclusions need to stay within the tested conditions.

If the goal is studying aroma loss, the grinding time should be fixed relative to the roasting time. A sample roasted longer before packing already has a different transformation history even at the same roast level. Similarly, if one group is ground before storage and another after, the difference in surface area creates a major source of error. The experiment's scope must clearly define product state from the very start.

References

• Effect of Temperature and Storage on Coffee's Volatile Compound Profile and Sensory Characteristics (2024), Foods.
• Impact of roasting and storage conditions on the shelf stability of Thai Arabica coffee (2025), Journal of Agriculture and Food Research 22, 102060, DOI 10.1016/j.jafr.2025.102060.
• Changes in volatile compounds and overall aroma profile during storage of coffee brews at 4 and 25 °C (2008), J. Agric. Food Chem., DOI 10.1021/jf703731x.
• Modeling of Greek coffee aroma loss during storage at different temperatures and water activities (2011), Procedia Food Science 1, 1111–1117, DOI 10.1016/j.profoo.2011.09.166.