Caffeine is at times the most divisive but least understood topic in coffee. I’m frequently asked how roasting temperatures, brewing methods, and even things like the country of origin affect a coffee’s caffeine content. A lot of beliefs fall somewhere between speculation and junk science.
The truth is that roasting doesn’t affect caffeine content in any dramatic way. Even then, brewing practices are far too inconsistent to communicate anything meaningful to an average consumer.
Almost all coffee is caffeinated. Caffeine is a stimulant alkaloid occurring naturally in dozens of species of plants, acting primarily as a defense against certain insects. Different species of coffee exist, and some have do have a high or low caffeine content in comparison with one another. Most coffee belongs either to the prized Coffea arabica or the hardier, but generally lower quality, Coffea canephora (robusta). Arabica, comprising almost all specialty coffee, contains in each seed (“bean,” although that’s a misnomer) about 1-1.5% caffeine. Robusta’s caffeine content is roughly double that.
Some varieties are naturally low in caffeine, such as the Laurina variety of Arabica, but so far, none of these low-caffeine varieties are widely available commercially. If a coffee is decaffeinated, that happens post-harvest, usually in a processing facility far away from the country of origin. Decaffeination processes use a solvent such as water to remove everything soluble from a batch of unroasted “green” coffee. The caffeine is removed from that slurry, then the decaffeinated slurry is reintroduced to the seeds (or the cellulose structures that were once seeds).
Coffee doesn’t undergo a significant thermal decomposition before it reaches 460 degrees Fahrenheit, which means it isn’t significantly affected by typical roasting temperatures. Only the absolute darkest roasts every approach that temperature, and those that do can afford to spend very little time in that zone. Most coffees never will.
But roasting does affect other physical properties of coffee including its density. As coffee roasts, it loses moisture and weight in a predictable manner while its seeds expand in size. These changes are more pronounced in longer and hotter (“darker”) roasts.
So a batch’s caffeine content is not affected, but the size and weight of its individual seeds are. Unfortunately, this presents a brewing conundrum. Without distinguishing between volume (scoops/tablespoons) and weight (grams/ounces), a darker roast could mean less caffeine per scoop but more per unit of weight.
(Why? Let’s say for simple math that a coffee seed contains maybe 1% caffeine. If we prep a 50lb batch of coffee, that’s a net of 8 ounces (.5lb) of caffeine. Let’s roast that batch. We can expect it to expand in volume and to lose perhaps 15% of its original weight. Our roasted batch now needs a larger container, but only weighs 42.5lb. Because all of the caffeine survives the roasting process intact, our larger in volume but less heavy batch still contains its original 8 ounces of caffeine.)
But any predictability this offers gets thrown out the window when we introduce other variables, such as differences from coffee to coffee where the typical caffeine content might not be precisely known, or differences in the serving size of drinks from one cafe to another, or the exact dose of coffee in a given drink, and whether it is adhered to consistently. (These issues are compounded with cold brew, which is somewhat more tolerant of inconsistent brewing, and thus is often victim to hasty or lazy preparation.) Caffeine is highly, rapidly soluble in water, so regardless of the brewing method, it’s correct to assume all of the available caffeine makes its way into the cup.
In other words, it’s reasonable to say you can’t make predictions about caffeine content in the context of a shop’s resources and workflow. Just enjoy what you drink: it’s caffeinated.
*This is a revision of an earlier piece on the same subject.