Chemical reactions require sufficient energy to get started, which is why they don't start until you reach certain temperature ranges. They then run until you run out of reaction inputs. So more time in a temperature range means more of the reaction products that you can produce in that range.

Roasting is tricky because the products of earlier reactions become inputs to subsequent reactions. So more maillard products changes the inputs when you reach the caramellization stage.

http://legacy.sweetmarias.com/roast.carlstaub.html

Trigonelline: A Nitrogenous Base Found in Coffee

Trigonelline is 100% soluble in water and therefore will end up in the cup. Trigonelline is probably the most significant constituent contributing to excessive bitterness. At bean temperatures of 445 degrees F, approximately 85% of the trigonelline will be degraded. This bean temperature represents a moderately dark roast. For lighter roasts there will be more trigonelline, hence bitterness, but also less sugar carmelization. Caramelized sugar is less sweet in the cup than noncaramelized sugar, so when properly roasted these two constituents form an interesting compliment to each other. Trigonelline melts in it's pure crystalline form at 424 degrees F Degradation of trigonelline begins at approximately 378 degrees F.. The degradation of trigonelline is one of the key constituent control flags for determining the best reaction ratio.

from: https://www.coffeechemistry.com/chemist ... -in-coffee

During roasting trigonelline partially degrades to produce two important compounds - pyridines and nicotinic acid - such that a very dark roast will only a fraction of its original trigonelline content....

...Nicotinic acid or also known was vitamin B3 (niacin) is produced by the demethylation of trigonelline at temperatures above 160°C-230°C where approximately 85% is decomposed. But when compared to green coffee - there is an overall net increase of about 10x from green to roasted.

According to lab experiments formation of nicotinic acid depends more on the roasting temperature than on the actual duration of the roast.

So it appears the balancing act we do is partially, if not significantly, related to the simultaneous degradation of trigonelline (bitter) and sucrose (sweet). During the roasting process coffee is getting less bitter at the same time it is getting less sweet. Bitter is mostly gone by 445*, but so is sweetness.

And the 2nd article indicates that the degradation or trigonelline seems more temperature dependent rather than time.

I'm in the camp that some bitterness is required for a satisfying cup of coffee. Controlling the ratio of bitter to sweet would be the goal for me.

I've found that in all my 'magic' roasts, final drop temps were on the low side - <410*. I wonder if the melting point of trigonelline at 424* is in play.

My brain hurts.

It depends greatly on the shape of the curve that creates the extended middle section of the roast and how much energy is retained going into the crack.

I would say the shape of the curve would be the same (declining RoR), but just elongated. The beginning temperature (dry end) and end temperature 1C would remain the same.

I think I'm beginning to answer my question. I have 50# pounds of an OK Colombia that has been replaced in my offering by a killer one. So I have been using it to learn a bit.

I did a few roasts last night, and while I haven't answered the OP question, I think I have answered the time vs temp question.

This first roast:
11:45 total roast
4:06 ramp
2:54 dev
425* final temp
Result: fully developed roast. Very chocolaty, not roasty in the least



Next roast:

12:42 total roast
5:36 ramp
3:06 dev'
408* final temp
Result: underdeveloped. vegetal notes,




Although the second roast ran longer in both ramp and dev phases, the final temp was much lower and resulted in an underdeveloped roast.

So it appears that specific temperatures above 408* need to be achieved to finish the roast, not just roasting longer.

Again, just thinking out loud.

Today I will try and answer my original question and do a 10 minute and a 12 minute roast, keeping everything except ramp time, including final temp, as identical as possible.

Almico wrote:I think I'm beginning to answer my question. I have 50# pounds of an OK Colombia that has been replaced in my offering by a killer one. So I have been using it to learn a bit.

I could be wrong and I certainly don't want to spark any kind of an argument, but I'm thinking that either of those roasts could have been successful depending on the specific bean.

I've long been dissatisfied with the idea that Maillard reactions and caramelization occur uniformly throughout the bean at specific sequential bean temperatures. It seems to make more sense recently to see my roasting as a cascade of reactions progressing from the outer to the inner as heat transfer drives the roast.

This would imply that Maillard reactions occur (at least briefly) within the ruptured endosperm after the release of water as steam. Similarly, there must be a degree of internal caramelization that would proceed very rapidly at the elevated temperatures at end of the roast.

This idea has proved useful for choosing the end of any given sample roast by scent. At the end of my standard declining rate of rise sample/exploratory roast I'm usually picking up a sweet scent in the trier during or immedieatly after first crack that I believe indicates Maillard reaction in the center of the bean. Within 10-15 seconds, this transitions to an acrid scent I believe is associated with caramelization and/or pyrolysis. For light (lighter, lightish) roasts, I'm dropping about 15 seconds later. When I do this, my coffee is more often than not completely developed and reasonably flavorful. Much beyond this and I'm usually moving into roast flavor in the cup.

I've been told that there are specific chemical compounds associated with both the sweet and acrid scent profiles, but this is all faith based science to me and I don't have enough faith in my understanding to quote what I've heard. All I know is that my roasting progressed from the "blind squirrel finding an acorn" stage to the "not automatically ruining everything I put in the drum" stage and now seems close to the "regularly managing to get a pretty good roast out of any reasonably good green" stage. -unless I take a phone call during the roast.

millcityroasters wrote:I've been told that there are specific chemical compounds associated with both the sweet and acrid scent profiles, but this is all faith based science to me and I don't have enough faith in my understanding to quote what I've heard. All I know is that my roasting progressed from the "blind squirrel finding an acorn" stage to the "not automatically ruining everything I put in the drum" stage and now seems close to the "regularly managing to get a pretty good roast out of any reasonably good green" stage. -unless I take a phone call during the roast.

That's too funny Steve. I think I'm at the same stage. I wish I had the option of smelling my roasts as they progress. With the exhaust hood on top of the roasting hopper, everything gets sucked out pretty efficiently. Unfortunately it is not easy to fashion a trier to accomplish the task. I'm working on making a 'sniffer' tube that I can place in the airflow to get a few whiffs at critical moments.

But like you say, a lot of the critical stuff happens very quickly. It's really uncanny how much ineffective science there is to roasting coffee. There is a lot of theory about what occurs, but very little in the way of full proof recipes to achieve best results. The proof is just how much bad coffee there still is all over the place. It's still more of an art form than a science.

I can't help thinking there is much more that can be achieved, but I think the equipment needs to evolve to handle far more precision, at least at this level.

millcityroasters wrote:I've long been dissatisfied with the idea that Maillard reactions and caramelization occur uniformly throughout the bean at specific sequential bean temperatures.

That this was ever an idea taken seriously surprises me, it shows how far the field has come.

millcityroasters wrote: It seems to make more sense recently to see my roasting as a cascade of reactions progressing from the outer to the inner as heat transfer drives the roast.

This summarises very well the information I've gleaned from serious scientific sources, thanks for articulating and promoting it!

Is this what Carl Staub addresses this with his "system transfer efficiency" comments here:

Best Reaction Ratio (BRR)

The best cup characteristic are produced when the ratio of the degradation of trigonelline to the derivation of Nicotinic Acid remains linear. The control model of this reaction ratio is a time/temperature/energy relationship. The environment temperature (ET) establishes the pyrolysis region for the desired chemical reactions while the energy value (BTU) and system transfer efficiency (STE) determines the rate of reaction propagation and linearity of Nicotinic Acid derivation to degradation of trigonelline.

Is his saying that the best reaction ratio, and the process of degradation of trigonelline to nicotinic acid, should occur linearly through the bean?

I assume some roasting machines and/or methods affect the transfer rate of heat through the bean. Do IR roasters transfer energy to the inner bean faster?

Hi Almico

Great with posting experiments

Almico wrote: So it appears that specific temperatures above 408* need to be achieved to finish the roast, not just roasting longer.
Again, just thinking out loud.

Or the conclusion could be that ROR matters - the rate of rise in temperature.

Its lower in your second roast. Both in the Maillard phase and the Development phase.

I have so many experiences that roasting to a certain Bean temperature does NOT give the same taste, so I don't pay attention to it. But I look at the speed of the roast, the ROR level.