Read Time: 15 minutes
“My goal is simple,” he once said. “It is complete understanding of the universe, why it is as it is and why it exists at all.” He spent much of his career searching for a way to reconcile Einstein’s theory of relativity with quantum physics and produce a “Theory of Everything.”
This is a quote from the now late Stephen Hawking and it seems terribly apropos to what I want to talk about today. I will never be on the intellectual level of Hawking, Einstein or any others of that magnitude. I would hope my desire to understand though is similar if a bit more narrow. With all due credit to Dr. Hawking, I would submit the following.
“My goal is simple. It is complete understanding of how to roast cocoa beans, why some roast one way and why others roast seemingly completely differently. In short I want a “Theory of Everything” pertaining to roasting all cocoa – A Grand Unified Theory”
Dissatisfied with oven roasting, both the results and trying to teach it, I have spent the last two years methodically putting together a system of drum roasting. It works beautifully for describing how to roast, stages the beans go through, the contribution of each phase and how to iterate the profiles to adjust the flavor profile just to your liking. It is sort of like the Theory of relativity. Unto itself it works well…except when it doesn’t. Like when you mix in quantum physics to Newtonian motion. In this case quantum physics is convection oven roasting and the nagging lack of explanations as to why it works for some beans and not others. Why does ‘long and low’ work at times when it is anathema to all I found true in drum profile roasting?
Well, I think I might have a piece of that puzzle, a piece of that theory of everything that links the two methods without contradicting either one. Right now it is based on just a handful of roasting observations but it seems to be self-consistent. Only time will tell if it holds true universally.
It all started with failure, as many discoveries do. I was preparing a matrix of roasts (24 samples in all) for the upcoming Roasting Seminar. I roasted 6 different beans, pulling kilogram samples at 220 F, 232 F, 245 F and whatever the given bean’s EOR was. The goal was to demonstrate how the flavors change over the course of the profile. All was going fine with Madagascar, Honduras and Nigeria. There was a steady decrease of astringency and bitterness, and an increase in fruit (where it existed) and chocolate as the profile progressed. Generally speaking I rated the final sample at EOR level my favorite (no surprise there).
Then the unexpected happened. I was tasting the Costa Rica Azul and suddenly nothing made sense. I thought I had mixed up the samples. My notes clearly indicated I liked the sample at 220 F the best. That made no sense at all. It was lacking in astringency, bitterness was low and there was a touch of sweetness. Although 20 years in the lab as a chemist has made me meticulous with my labeling schema, I thought maybe I had mixed up the sub-samples. I went back and pulled new samples from my stock and repeated my tasting…..only to get the same results.
So, instead of biasing my data to meet my expectations I dug in and tried to suss out what in the world would explain it. I’ll not bore you with any more of my internal dialogue but lay out what I think is going on.
There seems to be two types (at least) of bitterness and astringency (b/a). And very possibly two types of all the flavors - inherent and developed.
For some time now I have mentally railed against roasting ‘to preserve the flavor naturally present in the bean’ and that idea that roasting covers up those flavors. And really I still do, but there is also nuance there. Part of the nuance is that there ARE flavors present in the raw beans, but in my experience well over 90% of the time they are not flavors I want to preserve. Most notably there is a waxy, green astringency and metallic bitterness that decreases as you roast the beans. These are inherent flavors. They exist in the fermented, raw cocoa bean.
As you roast, both of these inherent flavors drop off (or are covered up by other developed flavors) but then start to increase in many beans as different bitter and astringent compounds are produced, usually when the beans have been subjected to too much heat too quickly or taken beyond their optimal EOR.
Circling around to the Costa Rica Azul, what I discovered is that it was remarkably low in inherent bitterness and astringency (b/a from here on out). That green/waxy flavor was not present. The result of this is that one of the reasons to roast past 232 F and into the 240s and 250s has gone away. At the same time, there appears (in this bean) to be little fruit and chocolate produced by the roast. A little is formed once you begin the Development phase (212-232 F) so again, there is little need to go into the finishing phase as there is little to finish, as it were. The only thing you get in the 232-250s is developed bitterness that over shadows that light, inherent fruit flavor, and light developed chocolate flavor.
From the best I can tell, and to summarize, this is what is occurring. There is low inherent bitterness, astringency and fruit. As the beans exits the drying phase at 212 F, a bit of chocolate and nut flavor is produced. At this point, there is no more you can do. You stretch out the Development phase time to assure water is gone and you have good heat penetration and full flavor development of what is possible. You don’t continue on up to 232 F as you both don’t need to and in doing so you risk developing other less pleasant competing flavors. Stopping at 220-225 F, but taking 3-5 minutes to do so, results in a delicate, nuanced flavored chocolate without any overshadowing of raw astringency since it didn’t exist in the first place.
The exciting part about all this is it is fully consistent with the previous Profile Roasting methodology in that the aroma of the roast tells you all of this is happening. In beans of this sort, there is a sharp prickliness that I associate with either too high of a ramp or EOR that occurs when the Development ramp is greater than 6-7 F/min or the beans reach 230 F.
This theory of inherent and developed flavors bridges the two big methods of roasting in my mind without contradicting either.
It explains why some beans work ‘long and low’ (low inherent b/a) and others don’t (higher inherent b/a). Costa Rica Azul (and up and coming Nahua) are the poster child for this.
It explains why some Criollo and certain Trinatario is often long and low and Forastero rarely works that way (low vs higher developed chocolate). Fiji is a great example of the later.
It explains why beans that roast great with a standard profile can’t be always be roasted long and low. Honduras is a fine example. It has a chocolate flavor needs to be developed to cover/mask/transmute the inherent b/a. Roast it only to 220 F and it will be very astringent.
So how do you use this? The first thing I want to remind you of is that the great majority of beans out there are NOT special snowflakes that require this treatment. It may appear there are lots of them, but I think it is more illusion than anything else. Cocoa beans selection seems to be going through what I would consider unnatural selection. Lots of people are roasting in convection ovens and the prevailing trend is long and low because otherwise you risk roast defects or they just don’t develop properly. The result is that people are using beans that only taste good with this form of treatment, self selecting and biasing the proportion of beans that can be roasted that way. Mostly I am trying to say don’t get sucked into thinking you need to hyper test every bean for its ability and/or need to roast low.
That said, should you find that a bean is going bitter and/or astringent at a much lower temperature than you expect (while drum roasting) then I suggest you peel and taste the raw beans.
If you find it is distinctly lacking in b/a then you could have a candidate for an EOR 220-230 F. Basically, long and low.
What does long and low mean? In a drum roaster it is different from convection ovens. In an oven, it often means 40-60 minutes (or more) at 220-250 F (or less). In Profile roasting in a drum it means still a 15-20 minute roast but you eliminate the Finishing phase.
Try your roast again and pay very special attention to the aromas coming off. Watch for sharp and prickly aromas once you are out of the Drying phase, and slow the roast down. Most beans can handle profiles where the Drying phase is 10 F/min, and 8-10 F/min in the Development phase. Because you still want a bit of time in the Development phase (2-4 minutes) then that means slowing down the ramps. Part of your Drying phase can still be 10 F/min, but I would recommend slowing the roast down around 180-190 F so you are maintaining control of the roast. Sort of slowing down in a car when you know a sharp turn is coming up. By the time you hit 212 F try to be in the 6-8 F/min range, and slow further to 4-6 F/min as you finish up your roast.
That is how I roast in my Royal. In the Behmor, I make sure I have a good full drum. 2.5 lbs great. It will naturally retard the roast. Around 212 F, the aroma coming off will change from basically nothing to something. Ambiguous I know, but will happen at 14-16 minutes. At that point, drop it down to P3/60% power to coast into the end. What is the end? Use the Force Luke. Well, maybe not that, but aroma will be your guide. 2-3 minutes, maybe 4-5 and that sharper prickly smell.
I know that is quite a few numbers. The big thing to remember is that if you are already roasting by aroma benchmarks, then nothing has changed except they are occurring earlier. Just continue to trust your senses.
But I will reiterate, don’t start off with this method. Only go to it when the tried and true roasting levels and profiles are not working as otherwise you can get tricked into under roasting something awful. And if the raw beans are bitter and/or astringent, this isn’t the roasting method you want.
So, there, that is my current gut feelings on a roasting GUT (Grand Unified Theory).