Case Study 2 — Roast Profiles at Equator Coffees

A Coffee Roaster's Maillard Problem

This case study draws on interviews with coffee roasters at three small-batch coffee companies (one in Portland, one in Chicago, one in California) and is presented as a composite. The "Equator Coffees" name and the specific characters are fictional; the data and the chemistry are real.

The Setup

Equator Coffees is a small specialty coffee roaster in northern California. It buys green coffee beans from cooperative farms in Ethiopia, Guatemala, Colombia, and Indonesia. The head roaster, Inés Vargas, is responsible for designing and executing roast profiles for each new lot of green beans that comes through the door. Her work is a daily exercise in applied Maillard chemistry, even though, until recently, she would not have used those words.

A coffee roast happens in a drum roaster — a rotating cylinder heated by gas burners or electric heaters, holding 1 to 60 kilograms of beans depending on the model. The beans tumble inside the drum, exposed to hot air and contact with the hot drum walls. The roast profile — the trajectory of bean temperature versus time — determines almost everything about the resulting coffee's flavor.

A typical roast profile for a light roast looks like this:

  • 0:00 — Beans loaded into preheated drum at about 200°C drum temperature. Bean temperature drops initially (cold beans absorb heat).
  • 3:30 — Bean temperature has climbed to 100°C. The "drying phase." Free water is evaporating out of the green beans. Color is still pale green.
  • 5:30 — Bean temperature reaches 140°C. Maillard reactions begin to run noticeably. Color shifts toward yellow.
  • 7:30 — Bean temperature reaches 160°C. Maillard chemistry is in full swing. Color shifts toward tan.
  • 9:30 — Bean temperature reaches 180°C. Caramelization joins Maillard. Color shifts toward light brown. Aroma begins to develop strongly.
  • 10:30 — Bean temperature reaches 196°C. "First crack" — a popping sound as steam pressure inside the bean ruptures cell walls. This is a major flavor event; the bean is now in the "first roast" stage.
  • 11:00 — Bean temperature 200°C. The roaster pulls the beans for a light roast.
  • 12:00 — Bean temperature 210°C. Medium roast.
  • 13:00 — Bean temperature 220°C. "Second crack" — the cell walls fracture further, fats migrate to the surface (the bean develops oil sheen), and the roast is approaching dark.
  • 14:00 — Bean temperature 230°C. Dark roast / French roast / espresso roast.

The whole thing takes 11 to 15 minutes. The window between "underdeveloped" and "burnt" is, on average, about 3 minutes.

The Chemistry Inside the Drum

Inside the bean during this roast:

Drying phase (0–4 minutes): Water evaporates. Bean structure shrinks slightly. Maillard cannot run yet because internal temperature is still below 140°C.

Maillard phase (4–9 minutes): Free amino acids in the bean react with reducing sugars (sucrose has hydrolyzed by this point, providing glucose and fructose). Pyrazines, pyrroles, furans, and Strecker aldehydes form. The pleasant "brewing-coffee" aroma begins to develop. Color shifts from yellow to tan to light brown.

Caramelization phase (overlapping with Maillard): Sugars also undergo caramelization, adding sweet notes and contributing brown color.

First crack (10:30): Cell walls rupture. Volatile compounds that had been trapped inside the bean are now able to escape, and the aroma intensifies dramatically.

Development phase (after first crack): This is where Inés has the most control. The bean is now soft and pliable; oils are migrating; flavor compounds are being formed and decomposed simultaneously. A 30-second extension at this stage produces a substantially different cup than pulling it at the end of first crack.

Second crack and beyond: Pyrolysis is increasingly happening. Smoky, charred flavors develop. Sugars are being burned rather than caramelized. Maillard products are decomposing into simpler, harsher molecules. The bean's character is shifting from "developed coffee" to "carbonized coffee."

The Profile-Design Problem

When a new lot of green coffee arrives at Equator, Inés performs a series of test roasts. She has a small 200-gram sample roaster for this. She tries the new beans at several different roast profiles:

  • Profile A — Quick light roast: Aggressive heat, 9-minute roast, pulled at first crack +30 seconds.
  • Profile B — Standard medium: Moderate heat, 11-minute roast, pulled at first crack +1:30.
  • Profile C — Slow medium: Gentle heat, 14-minute roast, pulled at first crack +2:00.
  • Profile D — Medium-dark: Moderate heat, 12-minute roast, pulled near second crack.
  • Profile E — Dark: Aggressive heat, 13-minute roast, pulled at second crack +30 seconds.

She brews each as espresso and as filter (drip), and tastes them with her team. Then she chooses one (or a couple) for the lot.

The variables she's controlling: - Total time - Temperature at first crack - Time between first crack and pull - Final pull temperature

What she's optimizing for is the flavor profile — the balance of acidity, sweetness, body, bitterness, and specific flavor notes. Different green beans want different profiles. A delicate Ethiopian Yirgacheffe wants a quick light roast that preserves its floral, citrus character; a deep Sumatran Mandheling wants a medium-dark roast that brings out its earthy, full-bodied character.

The chemistry guides Inés's choices.

Profile Choice as Maillard Choice

Let's translate.

Profile A (light roast): Maillard is mostly in early-stage. Many delicate Strecker aldehydes (3-methylbutanal, 2-methylbutanal, methional) are present and dominant. Some pyrazines are forming but not yet abundant. Caramelization is minimal because sugar hasn't had time to deeply react. The cup tastes bright, acidic, with floral and fruity notes. Body is light. Bitterness is low. This is what you want for a coffee with delicate origin character.

Profile B (standard medium): Maillard is in mid-late stage. Pyrazines and pyrroles are abundant, contributing the "roasted, nutty" character. Strecker aldehydes are still present but balanced by polymerized melanoidins. Caramelization has progressed. The cup tastes balanced, with roasted character but still some origin character. Body is medium. Bitterness is moderate. This is the workhorse roast for many coffees.

Profile C (slow medium): Same final temperature as B but more time spent in the Maillard regime. Maillard products have had longer to polymerize and to react with each other. Strecker products may have partially decomposed. The cup tastes rounder, more uniform in character, with less origin distinction but a smoother, more harmonious profile. Body is medium-full. This profile favors comfort over distinction.

Profile D (medium-dark): Maillard is at peak; second crack is approaching. Pyrazines are dominant. Some Maillard products are beginning to decompose into simpler bitter compounds. Cell wall fracture has released oils to the surface. The cup tastes bold, with strong roasted character, less acidity, more body. The origin character is partly obscured by the roast character. This is good for espresso blends.

Profile E (dark): Maillard products have decomposed substantially; pyrolysis is dominant. Smoky, charred notes appear. Origin character is essentially absent. The cup tastes bitter, dark, smoky, with very full body. This is the "espresso roast" or "French roast" — a profile choice that is more about the cooking than about the origin.

You can see the trade-off. As Inés moves from A to E, she trades preservation of origin character for development of roast character. There is no objectively "right" answer. There is only the question of what she wants the cup to be.

A Bad Day at the Roaster

Inés told me about one day in 2023 when the gas pressure to her roaster fluctuated. The drum temperature, normally stable within ±2°C of target, was bouncing around by 10°C. She didn't notice for the first batch — a Brazilian washed coffee she'd been roasting for years — and pulled it at her usual time.

When she cupped it that afternoon, it was wrong. The coffee was simultaneously underdeveloped and over-roasted. The early Strecker products were less abundant than usual; the late Maillard products were more abundant than usual; and there were some smoky, slightly burnt notes that didn't belong.

What had happened: the gas pressure swings meant the bean temperature had hit the Maillard threshold late (because of cold patches when the burner cut out) but had then overshot at certain moments when the gas surged. The chemistry had not run cleanly. The bean had spent too much time in the Maillard early stages (giving incomplete development) and had also spent peaks of time near pyrolysis (giving smoky notes). Pulling at her usual time had captured both of these chemical signatures.

She fixed the gas pressure, did a controlled roast on a new sample, and the chemistry came out clean again.

The lesson: Maillard chemistry is not just about the final temperature and time. It is about the trajectory. The path through Maillard space — how the bean got to its endpoint — matters for what it tastes like at the end.

What This Means for the Home Brewer

If you've never thought about coffee roast profiles before, here are some things to take away:

  • Different roasts emphasize different chemistry. Light roasts preserve more of the bean's origin character; dark roasts emphasize roast-derived character. Neither is better; they are different.
  • The roast profile is a chemistry recipe. A skilled roaster is making careful chemical decisions every minute of the roast.
  • Storage matters too. Maillard products in roasted coffee continue to evolve in storage. Coffee tastes different one week off-roast than three weeks off-roast. Most specialty coffees are best between 7 and 30 days post-roast — long enough for the freshly-formed compounds to integrate, short enough that volatiles haven't degraded.
  • Brewing is a Maillard extraction. When you brew coffee, you're essentially dissolving Maillard products into hot water. The extraction is selective — different compounds dissolve at different rates — which is why the same beans brewed by drip versus espresso versus French press taste different.

⚖️ Analyze This

  1. The "first crack" event in coffee roasting (around 196°C) is a physical event (cell wall rupture from steam pressure) but it has chemical consequences. List the chemical consequences and explain why first crack is such a critical decision point for the roaster.

  2. Profile A (quick light roast) and Profile C (slow medium roast) might have the same final bean temperature. Why do they produce different-tasting coffee? Use the concept of "Maillard trajectory."

  3. A roaster wanting to develop a specific Strecker product (say, methional, the baked-potato compound that contributes to coffee aroma) — what conditions of the roast might favor or disfavor methional?

  4. Inés's roaster had bad gas pressure one day. The result was a coffee that tasted both underdeveloped and over-roasted. Explain how the same coffee can have BOTH characteristics simultaneously, using Maillard chemistry.

  5. If you were to brew the same lot of green beans by light, medium, and dark roast and taste each side-by-side, which family of compounds (early Strecker products, late Maillard polymers, pyrolysis products) would dominate each cup? Predict the flavor profile of each.

  6. Coffee continues to develop in flavor for several days after roasting and then begins to deteriorate. Using your understanding of Maillard chemistry and the volatility of small molecules, explain (a) why fresh-roasted coffee tastes different at day 1 vs. day 14, and (b) why old roasted coffee eventually tastes flat.