Coffee Roasting Series #6 - Phases of Roasting

Today we are discussing the 5th and final principle of roasting, which covers the different phases of the roast. Mastery of these phases and how they affect coffee and its flavor are vital for roasters looking to have control over their roasting.
  1. Phases

Simplicity is helpful here. There are potentially many phases of the roast that you could analyze individually to understand better what happens chemically to coffee. That is beyond the scope of this blog series, because more detail is usually more confusing. You would never know why your coffee tastes bad if you kept focusing on what you should change in what 30s interval of a 12-minute roast.


For our purposes, there are 3 main phases. The three phases are, as I call them, Drying Phase (DRY), Maillard Phase (MAI), and Development Phase (DEV).


  1. A) Drying Phase. We covered this pretty well in the basic principles section. To elaborate or reiterate only slightly, drying time is not exactly about drying, but about establishing momentum and building inner pressure. You'll know you have too much pressure (and are using too much heat for too long) if you experience tipping, craters, or scorched parts. (Note that washed beans generally have some discoloration that can be mistaken by new roasters for light scorching, but is not at all scorched.)


Even if you do not get defects, if you find your roast tasting too "clear," "transparent," or  "weak-bodied" (this is different from flat, dull coffee) for your taste, you can use a little less heat during DRY to decrease momentum. Increased momentum means a faster roast. A faster roast means a faster MAI. As we'll discuss, a faster MAI can mean less Flavorful Reactions occurring. This is obviously not as bad as a relative absence of those reactions, and does not necessarily mean flat, dull coffee.


Again, DRY time will be between 5-6min, give or take 30s. Achieving this dry time is a useful marker but not a perfect marker of inner bean pressure. Your probe only reads the outside of the bean. It is possible to heat the outside of the bean to color change (Dry End) without getting enough heat into the bean. But if you have this problem, there is a simple solution: If you experience tipping, you have too much inner bean pressure (use less heat, lower charge, or longer soak). If you experience flavorless coffee, then you have too little inner bean pressure (use higher charges, more heat, shorter or no soaking).


For beans that have less moisture in them (decaf, low density, low elevation, older beans, beans exposed to air and light), DRY must be done carefully to not remove all the moisture. You need that moisture for the Flavorful Reactions to occur. DRY time should be similar between different beans, within a minute. A fast DRY time can mean too much heat has removed too much moisture inside the bean.


  1. B) Maillard Phase. MAI occurs between Dry End and First Crack. This is where your flavor begins to develop. Chemical reactions, such as caramelization, begin immediately at the point that color changes. Therefore, it is important to pay attention and note this time. Applying the same or more heat after this point, or seeing an increase of ROR after this point, will kill the sensitive Flavorful Reactions. You must take care to reduce heat and momentum thru this phase. The only thing to avoid is losing so much momentum that you don't have any heat left in your system to get through First Crack. You may enter MAI at a ROR of 28 F/min. You may hit First Crack around 20. (These are rough estimates and not specific to all machines, softwares, probes, coffees, etc.). As we discussed with ROR curves, the important thing is avoiding sudden spikes in heat or ROR that are not mere quirks of your probe and software.


You can make minor adjustments to flavor during the MAI phase, assuming you have good inner bean pressure. Without this internal pressure, no adjustments to MAI will help you. A shorter MAI will trend towards clarity and sweetness (to the extreme, lack of body, lack of mouth-feel, and under-caramelization). A longer MAI will trend towards full-bodied coffee (to the extreme, lack of note clarity and sweetness, coffee taste lingering too long, and an overbearing mouth feel). So there will be a range of balances that you will find enjoyable. Typically MAI roast times, depending on your roaster, volume of beans, etc, will fall between 3-4 minutes, give or take 30s.


For beans that you know are dryer, because of age, exposure, or processing, useless less airflow during Maillard (as well as Dry) can help retain some of that moisture in the bean. Increased airflow wicks away moisture.


  1. C) Development Phase. Like the other phases, the name "development" is slightly misleading. This phase has much more to do with outer bean development than inner bean. Flavor development is also happening throughout MAI. But DEV focuses on the final development of the roast that leads to the roast being light, medium, dark, or beyond. After First Crack, you want to continue the trend of a declining (however gently or radically) ROR curve.


If you would like to roast light coffee, you may want to approach First Crack with a slightly lower ROR than usual. This may require greater build of momentum earlier in the roast. Your ROR curve will need to decline rapidly after First Crack, as well. This allows you to have enough time in DEV to create Flavorful Reactions without letting your bean get darker than you want. A light bean may be only 15-20 degrees hotter, according to the bean probe, than it was at First Crack. If you track DEV as a percentage of total roast time, it should fall between 15% and 20% for a very-light to light roast.


Note: some beans, like some washed Ethiopian coffees I’ve roasted, tend to decline in ROR rapidly at this point on their own. If you intervene by adding heat, you risk increasing momentum and creating baked flavors by ruining Flavorful Reactions of this phase. You must know your bean through experience or research and anticipate the drop by supplying heat for maybe 20-30s about 1min before First Crack, then resuming normal heat settings. Be careful not to allow the ROR to increase before First Crack by doing this.


A word about "crashing" ROR: this is a major point online where people misunderstand popular advice, and lots of weird recommendations are made to try to use that advice. Or, people go through trouble to debunk what turns out to be a false interpretation of that advice.


ROR curves declining after First Crack is not a crash. If you are roasting light, that rapid decline is very necessary. If your probe, it's placement, or your software settings are set up in certain ways, you can always note a "flick and crash" around First Crack. This would be an anomaly, as discussed earlier in "ROR Curves."

At First Crack, the bean emits its own heat. That adds to the heat in the system and the bean probe reads this for you, so you see a short increase in ROR. Then, this short flash of heat is quickly drawn out as the beans crack open, and no longer becomes as significant of a source of heat. Your probe may read this is a sharp decline in heat present in the system.


But what happens to your probe is not exactly what happens to your bean. Your beans may not themselves be experiencing a sharp uptick in heat, followed by a sharp cooling, and then a rapid heating back up. If they did, this would potentially kill flavors. Be careful to note whether you are seeing flicks and crashes because something in your roasting method is leading to real temperature fluctuations in the system, or just because of anomalies with your roasting meter setup.


If you are roasting to medium, dark, or beyond, you will usually find you need to maintain more momentum during DEV to achieve darker color in an appropriate amount of time. Your ROR should still be declining gently over time, but not as rapidly. You also want to make sure you have adequate momentum going into Second Crack, as well. As you approach Second Crack, pressure is again building in the bean as they begin to produce more of their own heat. Therefore, a small increase in ROR just before Second Crack may be acceptable, although this shouldn't mean that you increase the heat yourself. You should likely be reducing heat, and increasing airflow here, unless you like to roast well beyond Second Crack.


Longer DEV will mean less: acidity, sweetness, and origin note flavor. You risk burnt coffee, overwhelming tobacco flavors, lack of acidity (acidity increasing perception of other flavors, sweetness, etc), and more. Shorter DEV will mean less caramelization, mouth feel, perceived body, smoothness. You risk a boring and unimpactful coffee.


DEV times can vary greatly, but roughly will be as follows: 1:45 for light roast, 2:30 for medium, 3+ for dark. This is dependent upon overall roast time, as momentum dictates how long to stay in each phase. It is more helpful to calculate DEV as a percentage of total roast time.


If you are roasting extra light to light, I recommend that you approach the roast with the intent to build greater momentum early in the roast, and have a slightly higher ROR reading than usual going into First Crack, before backing off heat for a rapidly declining ROR curve until drop. This is to ensure that you have enough inner bean pressure to create desirable Flavorful Reactions in a shorter DEV time, and avoid green or bland tasting coffee. This also builds internal pressure that continues to develop the coffee’s flavor over the next several days as it rests.


In my experience, lighter coffee can sometimes taste green for the first 3-7 days of cooling down. The flavor is still evolving. So don't throw it out if it's still green tasting on day 2. Give it a few more days. It might not be as good as it could have been, but it might not be underdeveloped trash either.