Coffee roasting; how complicated could that be? Just heat some beans until they turn brown, and you’re good to go. I recently found some local green coffee suppliers and realized I can buy small quantities for a fairly reasonable price. So I decided I would build my own coffee roasting setup. This is my journal.
I understand coffee roasting on a superficial level. Beans come dried and unroasted (“green”). By heating the beans, they undergo some sort of browning reaction that makes them delicious. Along the way, the beans “crack”, a lot like popcorn would. There are two cracks that happen, and passing through the “second crack” results in an oily dark roast.
With a bit of research I learned the browning process is due to a specific chemical process called the Maillard reaction. Louis-Camille Maillard was a French scientist who first characterized the process while attempting to reproduce biological protein synthesis. In essence, this is a heat-initiated reaction between amino acids and sugars and is ubiquitous throughout cooking. The browning on steak? Maillard. Brown crust on bread? Maillard. And most importantly, roasted coffee; Maillard.
Clearly the resulting compounds of the Maillard reaction are not identical in all of the above examples. During the reaction, a cascade of molecules are created which depend on the specific amino acids present. The resulting molecules are what give these foods their distinct flavors and aromas, however, the exact collection of molecules that are created are very hard to predict and control. For this reason, I am going to assume most coffee roasting discussion will be more of an “art” than a “science”.
[Is “art” what happens when the science is too complicated? The art of coffee roasting, the art of winemaking, cooking as an art; there are countless examples. Is this admitting that heuristic approaches are required because a process is too hard to fully control? By simplifying details, making approximations, and losing exactness, are we admitting science has failed to describe the process? No. What science has allowed us to do is make informed decisions about how and what to simplify to create a working description that matches the tools we have to enact the process.]
From what I have read, coffee gets tasty between 150-230 C (300-450 F). Along the way there are chemical events that happen at specific temperatures. For right now, those details are unimportant. What is important is that the roaster can reach temperatures up to 230 C (450 F), and will provide even heating to all of the beans. Most commercial solutions seem to involve either a rotating drum which agitates the beans as they heat, or blowing hot air such that the air agitates the beans.
I wonder if I can do better…
Further reading and sources: