The coffee drying phase is the first stage of a roast: it runs from the turning point, where the beans stop cooling and begin to climb, to the moment they shift from green to yellow. Its job is to drive off the bulk of the free moisture held in green coffee, because water has to leave before bean temperature can climb far enough for browning reactions to take hold.
Almost nothing that tastes like coffee is created here. What is created is the thermal foundation for everything after it — and roasters who chase flavour entirely in the last two minutes of a roast are often fixing, badly, a problem that was set up in the first four.
What the coffee drying phase actually does
Green coffee is a seed, not a dry ingredient. Specialty lots typically arrive at a roastery carrying somewhere in the region of 10–12% moisture by weight, with roasters commonly treating around 11% as a comfortable middle and anything much below 10% or above 12% as a coffee that will behave differently in the drum. Broader trade specifications are wider still — figures in the region of 8–12.5% are commonly cited as acceptable for storage and cup quality — and the number drifts with origin, processing, packaging and how long the lot has been sitting. That is exactly why two bags that look identical can want two different profiles.
That water is not simply an inconvenience to be removed as fast as possible. Roasters commonly describe moisture inside the bean as a heat-transfer medium: as it heats and turns to vapour it moves through the bean structure, carrying energy toward the core. A coffee that arrives unusually dry is often described as roasting flat or tasting hollow; a very wet coffee soaks up energy and can be sluggish and hard to steer. Either way, the drying phase is where the roaster meets that moisture head-on.
The phase begins at the turning point — the low point of the curve, where the cold charge stops pulling heat out of the drum and bean temperature starts to rise again. That page covers the mechanics of the turn itself; here, treat it simply as the starting gun.
Why the phase is largely endothermic
Evaporating water is expensive in energy terms. Turning liquid water into vapour at atmospheric pressure takes on the order of 2,260 kilojoules per kilogram — a large amount relative to what it takes simply to warm that water up. For most of the drying phase, a substantial share of the energy the roaster is pouring into the drum goes into that phase change rather than into raising bean temperature.
The practical consequence is a curve that climbs, but climbs steadily rather than steeply. The beans are absorbing heat, and the evaporating moisture is in effect cooling them from within, in the way sweat cools skin. Roasting literature generally describes the bean as net endothermic through this early period, tipping toward net exothermic only later, once browning and pyrolysis reactions begin generating heat of their own. Published estimates for where that crossover sits vary widely — sources differ, with figures from roughly 140 °C up to around 190 °C bean temperature appearing in different accounts — so it is best held as a direction of travel rather than a threshold to hit.
Because the phase is energy-hungry, energy withheld here is rarely recovered later.
Why the name is slightly misleading
Coffee does not finish drying at the end of the drying phase. Moisture continues to leave the bean right through browning and past first crack — the audible rupture that marks internal pressure finally exceeding what the bean structure can contain. A roasted bean still holds a small amount of water at drop.
What the phase actually covers is where the bulk of the free moisture goes: water sitting relatively loosely in the bean structure, as opposed to more tightly held water bound within the cellular matrix, which is more stubborn and leaves later and more slowly. Free moisture phase would be a more literal name. Drying phase stuck because it describes what dominates, not what exclusively happens.
This matters more than it sounds. Roasters who assume the bean is bone dry at yellowing tend to misread the behaviour of the curve afterwards, and blame the browning stage for momentum problems that are really residual moisture still asking for energy.
What a roaster actually smells and sees
Instruments tell you temperature and rate of change; your senses tell you what the coffee is doing. Through drying, experienced roasters lean on the nose at least as much as on the trier.
Aroma moves in a fairly reliable sequence. Early on the exhaust smells green and vegetal — grassy, hay-like, often described as drying straw or a freshly cut field. As the phase closes out, that gives way to something warmer and bready: baking biscuits, toast, cereal, a faint sweetness. Roasters commonly treat that grass-to-bread handover as a more trustworthy signal of the phase ending than colour alone, because it is less dependent on drum lighting and less prone to wishful thinking.
Colour moves from a dusty green through a briefly more vivid green as surface moisture flashes off, then to pale straw and finally a clear yellow. Trier samples across this stretch lose their damp, grassy density and start to feel drier and lighter in the hand. Silverskin begins loosening, and on many machines chaff starts appearing in the cyclone toward the end of the phase.
Be honest about the limits of colour. Roasters interpret yellow differently, and drum lighting, trier habits and eyesight all shift the call, which is one reason phase timings shared between roasteries are so often not comparable. Two competent operators on the same machine can log the same event half a minute apart.
Yellowing: the endpoint of the drying phase
The yellowing point is the marker roasters use to close drying and open browning. It is the moment a batch reads as unambiguously yellow rather than green, usually logged alongside a bean-temperature reading and a time.
Sources commonly place yellowing somewhere in the region of 145–160 °C (roughly 290–320 °F) bean temperature, but that number is close to meaningless between machines. Probe type, probe placement, batch size relative to drum capacity and how thoroughly the probe is bathed by the bean mass all move the reading, sometimes considerably. A yellowing temperature is a reliable reference for your own machine and a poor one for anyone else's.
What yellowing signals is that enough free moisture has left for bean temperature to climb more freely into the range where sugars and amino acids react in earnest. The chemistry that takes over from here — how browning builds colour, body and sweetness — belongs to the Maillard reaction and is worth reading separately. For drying-phase purposes, yellowing is simply the hand-off: the point at which the roaster's job changes from managing evaporation to managing browning.
Many roasters also treat yellowing as their most useful mid-roast checkpoint, because it is the first event in a roast that is both repeatable and early enough to act on. Arrive at yellow hot and fast and there is still time to moderate the run into first crack. Arrive late and short on energy and the options narrow quickly.
How long should the coffee drying phase be?
A guideline that circulates widely puts drying at somewhere around a third to a half of total roast time, with drying stretches often quoted at roughly four to seven minutes on commercial drum machines. Treat that as a description of common practice, not a target to engineer toward.
Fixed phase ratios are genuinely contested among roasters. Even the best-known ratio conventions in specialty roasting are presented by their own authors as rules of thumb that indicate how a roast progressed rather than as guarantees of anything, and published analyses of real batches routinely find plenty of well-regarded roasts sitting outside the recommended band. Plenty of respected roasters argue that chasing a percentage distorts the decisions that actually matter — how much energy the coffee carries into yellow, and whether the curve stays stable. A profile reverse-engineered from a target ratio can hit its numbers and still taste wrong.
The more useful framing is relational. Denser, higher-grown, wetter or larger-screen coffees generally ask for more energy and a longer drying stretch. Lower-density, drier or smaller beans need less. Batch size relative to drum capacity, machine type, airflow and even ambient conditions shift the answer again. What you are looking for is not a percentage but a smooth, decaying climb through the phase — the behaviour tracked by rate of rise, which is the instrument reading most roasters use to judge whether drying is progressing or stalling.
At a glance: the coffee drying phase
| Element | Detail |
|---|---|
| Starts at | The turning point — bean temperature stops falling and begins to climb |
| Ends at | Yellowing — the batch reads clearly yellow rather than green |
| What is happening | The bulk of the free moisture in green coffee evaporates; the bean absorbs heat and warms steadily; little meaningful flavour development yet |
| Energy behaviour | Largely endothermic; much of the applied heat goes into phase change rather than temperature rise |
| What you smell | Grassy, hay-like and vegetal early, shifting to bready, biscuity and toasty as the phase closes |
| What you see | Dusty green to vivid green to pale straw to yellow; trier samples feel drier; chaff starts to appear |
| Green moisture going in | Commonly around 10–12% by weight for specialty lots; varies with origin, processing and storage |
| Typical share of the roast | Commonly discussed as roughly a third to a half of total time — a widely used guideline, not a rule |
| What to watch | A smooth, steadily decaying rate of rise; enough momentum carried into yellow; no flat spots or spikes |
| Rushed it | Moisture still in the bean when browning starts; uneven core-to-surface development; sour, underdeveloped cups |
| Dragged it | Baked character — flat, papery, bready, with muted acidity and thin sweetness |
| Comparability between roasteries | Low: probe placement, machine type, batch size and colour judgement all move the numbers |
The levers you actually have
Only a handful of controls touch this phase, and what each does depends heavily on the machine in front of you.
- Charge temperature and batch size set how much stored energy is available before any gas is applied, and where the turn lands. A batch that half fills a drum behaves nothing like one that fills it.
- Applied heat is the main lever, but its effect is lagged: on most drum machines a gas change takes time to show up in the bean probe, so corrections made late in drying often land in browning instead.
- Airflow carries vapour and chaff out and shifts the balance between convective and conductive heat. Some roasters raise it through drying on wetter coffees; others hold it steady and change gas only. Both work, and the right answer is machine-specific.
- Drum speed, where adjustable, changes how the bed tumbles and how long each bean sits against hot metal — relevant mainly to scorching risk early on.
What goes wrong at each extreme
Rushed. Pushing hard and short through drying can leave moisture in the bean while browning is already underway at the surface. The outside runs ahead of the core, and the result is the classic underdeveloped profile: sharp, sour, grassy or peanutty notes, thin body, and a cup that tastes light no matter how dark the beans look. Scorching and tipping become likelier too, since aggressive early heat meets a bean that cannot yet absorb it evenly. The tell is a coffee that reads roasted to the eye and green to the palate.
Dragged. The opposite failure is the more insidious one, because the roast looks under control the whole way. Feed the batch too little energy and the curve flattens; the coffee spends a long stretch at a temperature warm enough to slowly cook it but not brisk enough to develop it. The cup goes baked: flat, papery, faintly bready, with the acidity sanded off and a sweetness that never arrives. Baked coffee is notoriously hard to diagnose blind, because nothing about it is obviously wrong — it is simply hollow. A drying phase that stalls, or that ends with the coffee limp on energy, is one of the most commonly cited routes there.
Between those extremes the target is unglamorous: a phase with enough heat to keep the climb decaying smoothly rather than lurching, ending at yellow with momentum still in hand.
How drying couples with the rest of the roast
No roast phase is independent, and drying is the one with the longest downstream shadow. Charge temperature and batch load set where the turn lands, which sets when drying begins; the energy carried out of drying then influences how briskly browning proceeds and how much control is left after first crack. A batch that reaches yellow late and cold tends to reach first crack late and cold as well, and the temptation to add heat at that point often produces exactly the ragged curve roasters are trying to avoid.
Honest limits apply throughout. Drum machines, fluid-bed machines and small home roasters distribute heat differently, so identical timings mean different things on each — a fluid-bed roaster driving heat almost entirely by convection will move through drying on a quite different clock from a heavy cast-iron drum. Probe placement changes every temperature quoted anywhere, including on this page. Green coffee moisture varies bag to bag, and roastery temperature and humidity move things again. Every number here is a starting point for your own record-keeping, not a specification.
The bottom line
The coffee drying phase is where a roast is set up rather than where it is decided, and yellowing is the marker that tells you the setup is finished. Get the energy right through this stretch and browning inherits a bean that is evenly heated and ready to react; get it wrong and every later adjustment is damage control. Log your yellowing time and temperature on your own machine, learn what grassy-to-bready smells like on your own exhaust, and treat every published percentage — including the third-to-a-half guideline — as a description of what other people's roasts happened to do, not a rule yours has to obey.