A comprehensive summary of roasting guides from Chocolate Alchemy and Dandelion Chocolate, focused on the science of what happens during roasting, with special attention to acetic acid removal, heat penetration, and flavor development.
The most common issue in craft chocolate is under-roasting — not in duration, but in heat penetration. Without monitoring actual bean temperature, roasters often stop when the surface seems done while the interior remains under-roasted.
"The current trend of not wanting to 'over roast' beans often results in people being terrified to fully roast, leading to under-roasted beans that are overly acidic, over astringent and lacking in good chocolate flavor (that is developed from a full roast)."
The Alchemist (John Nanci) divides roasting into three temperature-based phases. But they're not equal — only two of them significantly affect flavor.
"Think of it like tuning a guitar: development time is the string tension, slope is how fast you turn the pegs, and endpoint temperature is the note you're aiming for."
| Parameter | Target |
|---|---|
| Duration | 8-20 minutes (10-12 optimal) |
| Slope | 9-11°F/min |
What's happening: Not much chemically. Beans are normalizing moisture. This phase is about setting up momentum for the critical phases ahead.
Flavor impact: Minimal. An 8-minute vs 28-minute drying phase produced identical chocolate in blind tastings, as long as subsequent phases matched.
| Parameter | Target |
|---|---|
| Duration | 2-3 minutes (CRITICAL — do not exceed 5 min) |
| Slope | 8-10°F/min |
What's happening: Flavor precursors are built here — the working material that becomes actual flavors in the Finishing phase. Think of it like sautéing aromatics before simmering soup.
"Pushing this area of the profile does not in of itself give you flavor. It gives you working material that develops, in the Finishing phase, into flavors."
"I find adjusting the Development phase has the greatest impact on developing chocolate and fruit flavors."
How duration affects flavor:
| Duration | Slope | Effect |
|---|---|---|
| 2-2.5 min | 10°F/min (fast) | Accentuates fruit and chocolate; bright, dynamic |
| 2.5-3.5 min | 8°F/min (standard) | Balanced development; safe zone |
| 4-5 min | 5°F/min (slow) | Reduces acidity/astringency but flattens profile |
| >5 min | <4°F/min (too slow) | Bland, underdeveloped; lacks "pressure" |
Defects from bad Development:
- Too fast: Metallic taste, sharp tang (heat can't penetrate)
- Too slow: Flat, bland, underdeveloped (lack of "pressure" driving reactions)
| Parameter | Target |
|---|---|
| Duration | 3-5 minutes |
| Slope | 6-8°F/min (always ≤ Development slope) |
| EOR Temperature | 254-262°F for most beans |
What's happening: Chemical reactions are completed. This is also when excess acids are removed as heat penetrates to the bean core.
How slope/duration affects flavor:
| Slope | Effect |
|---|---|
| 8-10°F/min (fast) | Preserves more acidity; less heat penetration |
| 6-8°F/min (standard) | Balanced; acids driven off, flavors complete |
| 4-6°F/min (slow) | Maximum acid removal; deeper heat penetration |
How EOR temperature affects flavor:
| EOR | Character |
|---|---|
| 245-252°F | Bright fruit, higher acidity, origin-forward |
| 254-260°F | Balanced chocolate + dried fruit (sweet spot) |
| 262-270°F | Deep chocolate, caramel, diminished fruit |
| >270°F | Risk of burnt/metallic, origin character lost |
| Phase | Slope | Duration | Flavor Impact |
|---|---|---|---|
| Drying (→212°F) | 9-11°F/min | 10-14 min | None |
| Development (212→232°F) | 8-10°F/min | 2-3 min | Primary |
| Finishing (232→EOR) | 6-8°F/min | 3-5 min | Secondary |
| EOR | — | 254-262°F | Secondary |
- Acetic acid (vinegar) is created during fermentation when sugars convert to ethanol, then to acetic acid
- Acetic acid boils at 244.6°F
- BUT — you can smell acetic acid coming off well below that, often at 215-220°F
- Water vapor acts as a carrier, dragging acetic acid out of beans even below the boiling point
The surface temperature is NOT the interior temperature. This is the critical insight:
"Just because the surface of the bean is 250°F doesn't mean the interior is — unless you've held at 250°F for 10-15 minutes."
Without proper temperature monitoring or sufficient time, you may:
- Stop when the surface smells "done" while the interior is still under-roasted
- Not give enough time in the Finishing phase for heat to penetrate
- Not allow acids to migrate to the surface and volatilize
"A practical consideration is slowing your roast down near the end, giving the beans time to soak in the heat and time for the acetic acid to rise to the surface and evaporate."
Key targets for acid removal:
- Maintain 6-8°F/min slope in the Finishing phase (slower = more time for heat penetration)
- Hold for 3-5 minutes in the Finishing phase
- Target EOR of at least 254°F, preferably 256-262°F
- Look for the sharp/acrid aroma shift as the signal that acids are driven off
Heat Penetration: The Hidden Variable
Heat penetration is the overlooked factor in most roasting problems:
"When you profile roast in a drum roaster, where the Development phase is 2-4 minutes and the Finishing phase is 3-5 minutes, your cocoa will have sufficient heat penetration that the resulting chocolate will no longer have an astringency associated with raw cocoa."
- Beans need internal core temperature of 117-122°C (243-252°F)
- There's typically a 12°C (~22°F) gradient from surface to core
- At surface temps of 130-135°C (266-275°F), you get 97% degradation of astringent compounds
Implication: If your surface reads 250°F but you haven't given time for heat penetration, your bean interior could still be at 225-230°F — under-roasted.
"There is a sharp prickliness smell when you are pushing a bean too hard. It's the sign you are applying heat faster than it can soak into the bean and penetrate to center."
The Alchemist conducted a systematic tasting of Piura Blanco beans at 7 different EOR temperatures:
| EOR | Chocolate | Fruit | Character |
|---|---|---|---|
| 232°F | 2.0 | 2.0 | Sharp, sour, bile-like burn. Astringency 4/5. Under-roasted. |
| 245°F | 2.5 | — | Bright but shallow. Pith-like astringency. Not complex. |
| 252°F | 3.5 | High | Sweet spot begins. Passion fruit, apple blossom. Dynamic, bright. No astringency. |
| 255°F | 3.75 | Medium | Bitterness up but balanced. Fruit receding, chocolate advancing. |
| 260°F | 3.5 | Med-Low | Soft dried fruits. Jammy malic acid. Tannins add body. |
| 265°F | 3.0 | Low | Raisin, dried fruit dominant. Acidity flattening. Deeper, darker. |
| 270°F | 2.9 | Very Low | Crème brûlée, caramelized sugar. Sharp metallic edge. |
EOR Temperature → 232°F -------- 252°F -------- 270°F
| | |
Bright Fruit ████████░░░░░░░░░░░░░░░░░░░░░░░░
Fresh Acid ██████████░░░░░░░░░░░░░░░░░░░░░░
Astringency ████████████░░░░░░░░░░░░░░░░░░░░
Chocolate ░░░░░░░░████████████████░░░░░░░░
Dried Fruit ░░░░░░░░░░░░░░░░████████████████
Caramel/Roast ░░░░░░░░░░░░░░░░░░░░░░░░████████
| Level | EOR | Character |
|---|---|---|
| Under-Roasted | 232-245°F | Sharp, sour, acidic, high astringency. Unpleasant. |
| Light Roast | 245-252°F | Bright fruit, higher acidity, origin-forward. Requires extended conching. |
| Medium/Standard | 254-262°F | Balanced chocolate + dried fruit. The "95% of beans" sweet spot. |
| Dark Roast | 265-275°F | Deep chocolate, caramel, diminished fruit. Risk of metallic. |
| Over-Roasted | >275°F | Burnt, flat, metallic. Origin character destroyed. |
"If there is low chocolate flavor and low fruit flavors, you want to keep the EOR on the lower side (250°F and under). If you roast above that temperature then the bitterness that would usually be covered by chocolate and/or fruit becomes accentuated."
"Beans with fruit present and huge chocolate potential benefit from pushing the roast harder."
Translation:
- High-quality, fruity beans: Can handle higher EOR — developed chocolate masks bitterness
- Lower-quality or neutral beans: Keep EOR lower — nothing to cover the bitterness
10/8/6 @ 256°F
- 10°F/min from start to 212°F
- 8°F/min from 212°F to 232°F
- 6°F/min from 232°F to EOR
- 256°F EOR
12:00/14:30/19:00 @ 256°F
- 12 min to reach 212°F
- 14:30 to reach 232°F (2.5 min Development)
- 19:00 to reach EOR (4.5 min Finishing)
- 256°F EOR
Even with a temperature probe, aroma remains critical:
| Temperature | Aroma |
|---|---|
| 150-180°F | Unpleasant acids/ketones |
| 180-195°F | Wheat, bread, grain, hay |
| 212-220°F | Nuts, graham crackers, chocolate starting |
| >232°F | Nut smell fades, may get new acidic wave |
| 240-250°F | Fruit odors appear |
| ~260°F | Dip in acidic smell — potential pull time |
"Just because you are smelling something while roasting does not mean you are driving it off and out of your chocolate. Instead it means you have actually developed the aroma compound."
When you smell chocolate/brownie aromas, you're NOT losing those flavors — only trace molecular amounts escape. Continue roasting 2-3 more minutes after smelling chocolate.
When to stop: Wait until the aroma turns sharp/acrid in the 3-5 minute Finishing window, then stop. If it never turns sharp, stop at 262°F or 5 minutes in Finishing, whichever comes first.
Once you can reliably hit the standard profile, experiment with the levers:
| If you want... | Lever | Adjustment |
|---|---|---|
| More bright fruit | EOR | Lower (250-254°F), extend conching |
| More chocolate depth | EOR | Higher (260-265°F) |
| More dried fruit/raisin | EOR | Higher (262-268°F) |
| More caramel notes | EOR | Higher (265-270°F), risk metallic |
| Preserve delicate florals | Finishing | Lower EOR, slower slope (4-6°F/min) |
| Accentuate fruit + chocolate | Development | Faster (2-2.5 min at 10°F/min) |
| Reduce astringency | Development | Slower (3.5-4 min at 6°F/min) |
| Maximum acid removal | Finishing | Extend to 5 min, slow to 5°F/min |
Unlike coffee roasting where Maillard reactions are dominant, in cacao:
- Fermentation consumes most free sugars (reducing sugars needed for Maillard)
- Roasting temperatures (~270°F surface for 2-3 min) are too low for extensive Maillard reactions, which need "heat above 300°F for extended amounts of time"
- Strecker degradation requires Maillard byproducts, so it's also limited
Practical implication: You're primarily driving off acids and developing/completing flavor precursor reactions, not creating the deep Maillard browning that happens in coffee or bread.
- Load: 2-2.5 lbs (never less than 2 lbs — smaller amounts over-roast)
- Start on P1
- P1-P5 buttons become power controls during roast:
- P1 = 0%, P2 = 25%, P3 = 50%, P4 = 75%, P5 = 100%
- With a thermocouple modification, you can see actual bean temperature and hit proper targets
- The Behmor naturally achieves good slopes for cacao with proper loading
- Safety: At 4:30 remaining, press Start within 30 seconds or get Err 7
The Behmor's built-in 12-minute cooling cycle is adequate, but research suggests faster cooling (<5 min to 100°F) preserves 85-90% of volatile flavor compounds, while slow cooling (15-20 min) can lose 30-40%.
Simple improvement: Crack the door slightly after hitting Cool. This vents hot air faster while still allowing the fan to circulate. The fan is designed to pull air through the chamber, so leaving it slightly cracked (not wide open) throughout the cycle should speed cooling without disrupting airflow too much.
Important: Don't go crazy with rapid cooling (shop vacs, ice baths). Hyper-fast cooling can trap moisture inside the beans, causing viscosity issues in your chocolate. The goal is 3-5 minutes to get below 200°F, not 30 seconds to room temp.
| Defect Type | Symptoms | Cause |
|---|---|---|
| Under-Roasted | Acidic, astringent, sharp, sour, lacking chocolate | EOR too low, insufficient heat penetration |
| Over-Roasted | Burnt bitterness, flat, one-dimensional | EOR too high (>275°F) |
| Too-Fast Ramp | Metallic, sharp tang, surface scorched/interior raw | Slope too steep, heat can't penetrate |
| Too-Slow Development | Bland, flat, underdeveloped | Development >5 min, lack of "pressure" |
"It is actually REALLY hard to over-roast cocoa — I'm talking you have to try to over roast and you still might fail."
If you have batches that are too acidic:
"By the natural process of grinding and aerating, a lot of volatiles (acetic acid included) will be driven off. Raising the temperature a little (like pointing a heat lamp on it) will help. Adding a fan to blow gently over the top of the chocolate can also do wonders."
- Refine at 135-145°F
- Allow 24-48 hours
- Gentle airflow helps volatilize acids
- Make milk chocolate — milk proteins convert acidic compounds into "softer aldehydes and ketones"
- Use for cooking/baking where acidity is masked
| Symptom | Cause |
|---|---|
| Sour/vinegary | Acetic acid not driven off |
| Astringent/puckering | Raw cocoa compounds not degraded |
| Sharp or harsh | Unpleasant acids remain |
| Lacking chocolate flavor | Flavor precursors not fully developed |
| "Green" or raw taste | Insufficient heat penetration |
| Thin body | Reactions incomplete |
If your chocolate has these characteristics, you likely under-roasted (internal temp ~232-245°F). You're in the worst zone: all the negatives (acid, astringency) without the positives (fruit brightness OR chocolate depth).
- Two flavor levers: Development duration (primary), Finishing duration/EOR (secondary). Drying doesn't matter.
- Monitor bean temperature, not air/oven temperature
- Hit proper phase targets: 212°F → 232°F → 254-262°F EOR
- Maintain correct slopes: ~10/8/6 °F/min through each phase
- Allow adequate Finishing time (3-5 min) for heat penetration and acid removal
- Use aroma as confirmation, not primary guide
- Don't fear roasting — under-roasting is far more common than over-roasting
- The sharp/acrid aroma shift signals roast completion
- Ask the Alchemist #183 - Acetic Acid
- Ask the Alchemist #200 - Profile Roasting Intro
- Ask the Alchemist #201 - Drying Phase
- Ask the Alchemist #203 - Development Phase
- Ask the Alchemist #204 - Course Correction
- Ask the Alchemist #205 - Finishing Phase
- Ask the Alchemist #206 - Maillard Reactions
- Ask the Alchemist #214 - When to Stop
- Ask the Alchemist #235 - Under-roasting Problems
- Ask the Alchemist #278 - Roasting Profiles
- Ask the Alchemist #289 - Slope and RoR
- Ask the Alchemist #307 - Heat Penetration
- Ask the Alchemist #341 - Profile Notation
- Behmor Roaster Guide
- Profile Roasting Seminar
- Complete Chocolate Making Guide
Dandelion Chocolate, a renowned San Francisco bean-to-bar maker, takes a notably different approach to roasting. Understanding both philosophies illuminates the trade-offs in roasting decisions.
"We prefer a very light roast, which doesn't work on all beans."
"We like to roast them long and slow, retaining as much flavor as we can."
Key details:
- Roast 5 kg batches in a modified coffee roaster
- Do 8-16 test roasts per origin before finalizing a profile
- Focus on preserving bright, origin-specific flavors
- Only use two ingredients: cacao and sugar (no vanilla, lecithin, or added cocoa butter)
Dandelion doesn't publish specific temperatures or times, but using Chocolate Alchemy's framework:
- Lower EOR temperature: Perhaps 240-250°F instead of 254-262°F
- Shorter Finishing phase: Stopping before the sharp/acrid aroma shift
- Preserving more acidity: Intentionally retaining bright, fruity acids
- Lower slope/RoR: Perhaps 4-6°F/min instead of 8-10°F/min in Development
- Extended total time: Maybe 25-35 minutes total instead of 18-22 minutes
- Gentler heat application: Lower temperature differential
This approach allows more even heat penetration without high surface temps, and preserves delicate origin aromatics.
Dandelion compensates with professional equipment:
- Industrial melangers running 48-72+ hours (vs 12-24 hours for home makers)
- Dedicated conching equipment with controlled airflow
- Precise temperature control during refining
"When roasted lightly, our Madagascar beans have the bright flavors we think taste amazing as a chocolate bar (after they've been conched and sugar has been added) but we don't necessarily love those flavors as much as a stand-alone experience."
For their Whole Roasted Beans (eaten without sugar), they roast longer:
"By increasing the roast time, we skim off some of the sharper acidic flavors, leaving a nutty, citrusy flavor profile that is more snackable."
Where Chocolate Alchemy emphasizes driving off volatiles during roasting, Dandelion relies on refining/conching:
"The friction from the rollers creates heat, which causes some of the harsher flavors to boil off."
"The longer a chocolate is in a melanger, the mellower it tends to be."
| Aspect | Chocolate Alchemy | Dandelion Chocolate |
|---|---|---|
| EOR | 254-262°F | ~240-250°F (speculative) |
| Development slope | 8-10°F/min | ~4-6°F/min (speculative) |
| Total time | ~18-22 minutes | ~25-35+ minutes |
| Acid removal | Primarily in roaster | Primarily in conching |
| Target audience | Home/small batch | Professional production |
| Conching time | 12-24 hours | 48-72+ hours |
Think of it as a "budget" of acidity you need to remove:
- Dandelion approach: Remove 30% in roaster, 70% in conching
- Home maker approach: Remove 70% in roaster, 30% in conching
If you can't conche long enough, you'll end up with acidic chocolate.
For home makers with:
- Limited conching time (12-24 hours)
- Smaller melangers that generate less heat
- No dedicated conching equipment
...a fuller roast (per Chocolate Alchemy's guidelines) is more practical. "Light and slow" without professional conching equipment = acidic chocolate.
- Know your equipment limitations: If you can't conche for 48+ hours, roast more fully
- Light roasts preserve origin character but require more post-roast processing
- Acidity can be addressed at multiple stages: roasting, refining, or conching
- Test extensively: Dandelion does 8-16 test roasts per origin
Making Chocolate: From Bean to Bar to S'more
- 368 pages, Clarkson Potter
- Covers sourcing, fermentation, roasting, winnowing, refining, tempering
- 30 recipes from their pastry kitchen
A review of peer-reviewed studies on cacao roasting confirms the practical framework above and adds a few useful insights.
Volatile sulfur compounds (DMDS, DMTS) — responsible for smoky, meaty, savory notes — increase ~350x when bean temperature exceeds 135°C. This validates the Finishing phase as a distinct flavor zone, not just "more of the same."
García-Alamilla et al. (2017): Roasting above 135°C produced measurably different aroma compound profiles.
Optimal water activity (aw) of 0.3 maximizes pyrazine development. Pyrazines contribute ~40% of cocoa's characteristic aroma (nutty, roasty notes). This validates not rushing the Drying phase — proper moisture reduction sets up better flavor development later.
Academic studies confirm a 10-12°C (~18-22°F) gradient between bean surface and core during roasting. This is why:
- Heat penetration requires TIME, not just temperature
- Stopping based on surface temperature alone leads to under-roasted interiors
- Slower Finishing slopes (6-8°F/min) allow the core to catch up
Fermentation consumes most of the reducing sugars needed for Maillard reactions. This means:
- Cacao roasting is primarily about driving off acids and completing flavor precursor reactions
- You have a narrower "sweet spot" before going from developed → burnt
- Over-roasting quickly becomes unpleasant because there's less browning chemistry to buffer it
Bean pH (set by fermentation) affects which aroma compounds form:
- Lower pH beans: Produce more pyrazines (nutty/roasty)
- Higher pH beans: Produce more methylpyrazines (different roast character)
You can't change this during roasting, but it explains why identical profiles taste different across origins.
| Finding | What It Means |
|---|---|
| 275°F VSC threshold | Your thermocouple hitting 275°F is chemically significant, not arbitrary |
| Water activity 0.3 optimal | Don't blast through drying — proper moisture reduction helps flavor later |
| 22°F surface/core gradient | Your acidic roasts were likely a heat penetration problem — core never got hot enough |
| Limited Maillard potential | Explains the narrow sweet spot; over-roasting goes bad fast |
| pH affects aroma profile | Same roast profile will taste different on different origins |
Bottom line: The science confirms what Chocolate Alchemy teaches practically — monitor actual bean temperature, give the Finishing phase enough TIME (not just temperature), and don't fear pushing to 254-262°F.
- How We Make Chocolate
- Education Station: Whole Roasted Beans
- How Low Can You Go?
- Le Grande Experiment Series
- Making Chocolate Book
This section synthesizes findings from peer-reviewed studies on cocoa roasting chemistry, kinetics, and flavor development. The research strongly validates Chocolate Alchemy's practical recommendations while adding quantitative data and mechanistic understanding.
- Quelal et al. (2023): "Key Aromatic Volatile Compounds from Roasted Cocoa Beans, Cocoa Liquor, and Chocolate" - Comprehensive review of volatile compounds through processing chain (MDPI, open access)
- Rojas et al. (2022): "Kinetic Studies on Cocoa Roasting Including Volatile Characterization" - Thermochemical kinetics and volatilization windows (ACS Food Science & Technology)
- Rojas et al. (2022): "Physicochemical Phenomena in the Roasting of Cocoa" - Engineering review of roasting processes (Food Engineering Reviews)
- Frauendorfer & Schieberle (2008, 2019): Changes in key aroma compounds during roasting (Journal of Agricultural and Food Chemistry)
Volatile Sulfur Compounds (VSCs):
Research confirms that bean temperatures above 135C (275F) trigger dramatic chemical changes:
- Dimethyl disulfide (DMDS): Increases ~350x above 135C
- Dimethyl trisulfide (DMTS): Increases ~280x above 135C
- These create smoky, meaty, cabbage, gasoline-like off-notes
- They persist through chocolate-making and cannot be removed by conching
Smoky off-flavor thresholds:
- 3-ethylphenol: Acceptable <2 g/kg; exceeds threshold dramatically above 135C
- 2-methoxyphenol: Acceptable <70 g/kg; can reach 221 g/kg in over-roasted beans (OAV of 122)
Practical conclusion: The 254-262F EOR target ensures bean cores stay safely below 275F, avoiding VSC formation while still achieving proper chocolate development.
Temperature-dependent formation:
- 100C (212F): Pyrazines begin forming (minimal amounts)
- 110-121C (230-250F): Peak pyrazine formation rate
- >130C (266F): Formation continues but some simpler pyrazines start volatilizing
Types and their importance:
Simple pyrazines (2,5-dimethylpyrazine):
- High odor thresholds (129 g/kg)
- Give basic "roasted, biscuit" notes
- Form earlier in roasting
Complex pyrazines (2,3-diethyl-5-methylpyrazine, 2-ethyl-3,5-dimethylpyrazine):
- Much lower odor thresholds (0.5-7.2 g/kg)
- Contribute "chocolate, cocoa, earthy, baked potato" notes
- Require sustained heat (Development + Finishing phases)
- Odor Activity Values (OAV) of 6-40 in well-roasted beans
Validation: This explains why Development phase (212-232F) is the "primary flavor lever" - it's the window where flavor precursor reactions generate the compounds that become complex pyrazines in the Finishing phase.
The key compounds:
3-methylbutanal (from leucine):
- Dried beans: 60-85 g/kg
- Roasted beans (130C, 30 min): 721 g/kg (~9x increase!)
- Odor threshold: 5.4-80 g/kg
- OAV in cocoa liquor: 9-134 (dominates chocolate aroma)
- Character: "Malty, chocolate"
2-methylbutanal (from isoleucine):
- Similar profile to 3-methylbutanal
- OAV: 50-320 in roasted beans
- Character: "Malty, chocolate"
Phenylacetaldehyde (from phenylalanine):
- Dried beans: ~23 g/kg
- Roasted beans: 340-534 g/kg (15-23x increase)
- OAV: 10-245
- Character: "Honey, floral, sweet"
Formation timeline:
- Begin forming during Development (212-232F)
- Continue forming during conching (why under-roasted beans can't be fully "fixed")
- Require proper core temperature to generate sufficient precursors throughout bean
Critical insight: Surface roasting creates Strecker aldehydes only in outer layers. Proper heat penetration generates them throughout, creating fuller chocolate flavor.
Temperature-dependent removal rates:
From multiple studies on fermented dried beans:
| Roast Temperature | Duration | Acetic Acid Remaining | Reduction |
|---|---|---|---|
| Unroasted | - | 1,240-2,048 g/kg | Baseline |
| 95C (203F) | 30 min | ~2,048 g/kg | 0% |
| 125-130C (257-266F) | 30 min | 1,024-1,200 g/kg | 50% |
| 135-140C (275-284F) | 30 min | 512-600 g/kg | 70-75% |
Odor Activity Values (OAV):
- Dried beans: OAV = 154-467 (extremely noticeable - "sour, vinegar")
- Roasted at 130C: OAV = 49-82 (still very noticeable)
- Cocoa liquor: OAV = 16-48 (noticeable but balanced)
- Final chocolate: OAV = 13-30 (acceptable)
Target: Reduce acetic acid OAV to <30 in final chocolate. Roasting must achieve 60-70% reduction minimum.
The volatilization window:
Research using thermogravimetric analysis shows:
- Acetic acid begins volatilizing at ~215-220F (via steam distillation with departing water)
- Continues through ~302F (not just at 244.6F boiling point!)
- Most efficient removal: 250-280F (121-138C)
Why it takes time:
- Heat must penetrate to bean core (where most acetic acid resides)
-
- Acid must migrate from interior to surface via moisture/steam
-
- Surface must be hot enough for volatilization
Validation: This confirms why Chocolate Alchemy emphasizes 3-5 minutes in Finishing phase (232-262F) - you need dwell time for acid migration and escape.
Roasting vs. Conching efficiency:
| Process | Duration | Acetic Acid Removal | Other Volatile Acids |
|---|---|---|---|
| Roasting (130C) | 30 min | 50-60% | 60-75% |
| Conching (80C) | 6 hours | 20-30% | 10-15% |
| Conching (80C) | 48 hours | 50-60% | 40-50% |
Conclusion: Roasting is 2-3x more efficient at removing volatile acids per unit time. This validates doing acid removal primarily during roasting.
Key volatile esters (fruity notes):
| Compound | Character | Dried Beans OAV | Roasted 130C OAV | Loss |
|---|---|---|---|---|
| Isoamyl acetate | Banana, fruity | 91 | 40-60 | 40-50% |
| 2-heptanol | Citrus | 11 | 5-8 | 50-60% |
| 2-nonanone | Fruity, fresh | 31 | 15-20 | 40-50% |
| Ethyl phenylacetate | Fruity, honey | 1-3 | 0.5-2 | 50-70% |
Floral compounds:
| Compound | Character | Dried Beans OAV | Roasted 130C OAV | Loss |
|---|---|---|---|---|
| Linalool | Floral, sweet | 17 | 8-12 | 50% |
| 2-phenylethanol | Honey, rose | 13 | 8-10 | 30-40% |
Temperature sensitivity:
- 95-120C (203-248F): Minimal loss (10-20%)
- 125-135C (257-275F): Moderate loss (40-50%)
- >140C (>284F): Heavy loss (70-80%)
The Linalool/Benzaldehyde Ratio:
Proposed as fine flavor indicator:
- Fine grade cocoa: Ratio >0.3 (more floral notes preserved)
- Bulk grade cocoa: Ratio <0.3
- This ratio persists through roasting if you avoid over-roasting
Practical application:
- Fine Criollo/Nacional: Target 248-255F EOR to preserve fruit/floral (can tolerate slightly higher astringency, extended conching)
- Bulk Forastero: Target 254-262F EOR (less fruit to preserve anyway, priority is chocolate development and astringency reduction)
Critical finding: Optimal water activity (aw) for pyrazine formation = 0.3
Moisture progression through roasting:
| Stage | Moisture % | Water Activity | Pyrazine Formation |
|---|---|---|---|
| Dried beans | 6-7% | 0.4-0.5 | Not yet started |
| After Drying (212F) | 2-3% | ~0.3 | Optimal! |
| After roasting | 1-2% | 0.15-0.2 | Slowing down |
Why this matters:
- Too wet (aw >0.5): Competing hydrolysis reactions reduce pyrazine yield
- Optimal (aw = 0.3): Maximum Maillard reaction efficiency
- Too dry (aw <0.2): Reactions slow dramatically
Validation: This is why Chocolate Alchemy says "don't blast through the Drying phase" - proper 10-14 minute progression to 212F achieves the aw = 0.3 sweet spot right as you enter the Development phase where pyrazine precursors form.
Hydrophobic amino acid content (flavor precursors):
High-quality origins (Central America, Caribbean, Peru, Papua New Guinea, fine Criollo/Nacional):
- Hydrophobic amino acids: 68-73% of total
- Rich in leucine, phenylalanine, alanine
- High peptide content from fermentation
- Result: More flavor precursor potential
Bulk origins (West Africa, SE Asia, basic Forastero):
- Hydrophobic amino acids: 50-60% of total
- Lower peptide content
- Result: Limited flavor precursor potential
Practical implication:
Origin matters chemically, not just marketing:
- High-peptide beans develop more complex chocolate at same roast level
- Low-peptide beans may benefit from slightly higher EOR (260-262F) to maximize flavor from limited precursors
- You can't "create" flavor precursors during roasting - you're limited by what fermentation produced
Recent discovery (2018-2020 research):
Storing whole pods 3-7 days before fermentation:
- Reduces fermentable sugars (less acid formation)
- Maintains or slightly increases amino acids
- Creates 4:1 fructose:glucose ratio (vs 2:1 without storage)
- Results in 40-50% lower acetic acid after fermentation
Effect on roasting:
- Pod-stored beans can tolerate slightly lower EOR (250-255F)
- Less acid to remove = more flexibility
- Flavor profile: more fruity, less acidic
If sourcing: Ask about pod storage practices - these beans are more forgiving.
Convective (hot air):
- Most uniform heat distribution
- Faster moisture removal (can over-dry surfaces)
- Best for large batches (>10 kg)
- Can achieve target temps in 15-20 minutes
- Risk: Surface scorching if too fast
Conductive (drum) - like Behmor:
- Better heat penetration for small batches
- More gradual moisture removal (better aw control)
- Ideal for 1-5 kg batches
- Requires 18-25 minutes for proper heat soak
- Advantage: Time allows proper phase progression
Superheated steam:
- Very rapid heating
- Sugar degradation increases 40% vs conventional
- More pyrazines but also more burnt notes
- Not recommended for craft chocolate
Fluidized bed:
- Fastest pyrazine formation (excellent agitation)
- Can achieve full roast in 12-15 minutes
- Requires specialized equipment
- Risk: Uneven roasting if particle sizes vary
Validation: Small drum roasters are ideal for craft roasting - they provide time needed for heat penetration without over-drying surfaces.
Heat penetration rates:
- Small beans (<1.0g): Reach target temp ~30% faster
- Large beans (>1.5g): Require 4-6 minutes longer in Finishing
- Mixed sizes: 15C internal temperature variation
Fat content effects:
- Higher fat (54-56%): Slower heat penetration (fat acts as thermal buffer)
- Lower fat (48-52%): Faster heat penetration
Practical recommendation: Grade beans by size before roasting if possible. If mixed, extend Finishing phase by 2-3 minutes to ensure large beans reach target.
Critical finding:
- Rapid cooling (<5 min to 100F): Preserves 85-90% of volatile compounds
- Slow cooling (15-20 min): Loses 30-40% of desirable volatiles
- Very slow (>20 min): Continues Maillard reactions, can develop burnt notes
Mechanism: Volatiles continue escaping as long as beans stay above 150F. Chocolate-defining pyrazines and Strecker aldehydes are lipophilic but will volatilize from hot beans.
Practical tip: Use forced-air cooling immediately after pulling beans. A simple fan makes measurable difference in final flavor intensity.
Total acidity reduction needed: ~80-85%
Two valid approaches:
Route A (Chocolate Alchemy / Home approach):
- Roasting: 70% acid removal
- Conching (12-24 hrs): 15% acid removal
- Advantage: Removes acids when you WANT volatiles leaving
- Result: Cleaner, more chocolate-forward
Route B (Dandelion / Professional approach):
- Roasting: 30% acid removal (lighter roast)
- Conching (48-72 hrs): 55% acid removal
- Advantage: Preserves more fruit/floral notes
- Challenge: Requires extended conching time and equipment
Volatile changes during conching (80C, per hour):
- Acetic acid: -15%/hr (good!)
- Simple pyrazines: -8%/hr (acceptable)
- Complex pyrazines: -3%/hr (minimal)
- Strecker aldehydes: -2%/hr (minimal)
- Fruity esters: -12%/hr (significant loss!)
Conclusion: Chocolate Alchemy's fuller roast is more practical for home makers - you lose volatile acids during roasting (desired) rather than losing fruity notes along with acids during extended conching.
Criollo varieties:
- Higher linalool content, lower baseline acidity
- Optimal EOR: 248-255F
- Strategy: Preserve floral notes, tolerate slightly higher astringency
- Linalool/benzaldehyde ratio: Maintain >0.4
- Can use slower Finishing slope (4-6F/min) for complexity
Forastero varieties:
- Lower fruit, higher baseline acidity, more polyphenols
- Optimal EOR: 258-262F
- Strategy: Full astringency reduction priority (97% polyphenol degradation needs sustained >257F)
- Standard or fast Finishing slope (6-8F/min)
Nacional (Ecuador):
- High fruit potential, moderate acidity
- Optimal EOR: 252-258F
- Benefits from pod storage + moderate roast
Trinitario (hybrid):
- Variable (genetics depend on Criollo/Forastero ratio)
- Optimal EOR: 254-260F
- Test and adjust; monitor linalool/benzaldehyde ratio
CCN-51 (high-yield hybrid):
- Sweet/fruity when fresh, but bland chocolate potential
- Lower amino acid content
- Optimal EOR: 260-265F (push harder to develop chocolate from limited precursors)
For health-conscious chocolate makers:
| Bean Temperature | Epicatechin Loss | Flavor Development | Astringency |
|---|---|---|---|
| 245F | ~30% | Low | High |
| 250F | ~50% | Moderate | Moderate-High |
| 260F | ~70% | Good | Low |
| 270F | ~85% | Excellent | Very Low |
Decision points:
- Health-focused: Roast to 245-250F, accept astringency, conche longer
- Flavor-focused (most craft makers): 254-262F standard
- Deep chocolate: 265-270F, minimal flavanols remain
Note: If health is priority, consider alkalization (Dutch process) as alternative astringency reducer that's less temperature-dependent.
| Parameter | Light Roast | Standard Roast | Dark Roast |
|---|---|---|---|
| EOR Surface Temp | 245-252F | 254-262F | 265-275F |
| Core Temp (est.) | 240-248F | 250-258F | 260-270F |
| Acetic Acid Removal | 40-50% | 70-75% | 85-90% |
| Pyrazine OAV | 5-15 | 20-40 | 40-60 |
| Fruity Ester Retention | 70-80% | 40-50% | 10-20% |
| Floral Retention | 60-70% | 40-50% | 10-20% |
| VSC Formation | Minimal | Low | Moderate-High |
| Epicatechin Retention | 70-80% | 35-45% | 10-15% |
| Best For | Fine Criollo, extended conching | 95% of beans, balanced | Strong chocolate flavor |
Strongly validates:
- 254-262F EOR is chemically optimal (pyrazine formation complete, below VSC threshold)
- 3-5 minute Finishing phase necessary (acid migration and volatilization time)
- 6-8F/min Finishing slope (heat penetration vs. surface scorching balance)
- Development phase (212-232F) as primary flavor lever (precursor formation window)
- Time matters more than peak temperature for heat penetration
- Under-roasting more common than over-roasting
Adds quantitative detail:
- Specific OAV values showing which compounds dominate flavor
- Temperature thresholds for different chemical reactions (135C VSC threshold, aw 0.3 optimal)
- Kinetic data on compound formation and volatilization rates
- Origin-based amino acid profiles explaining why some beans develop better flavor
Research gaps still exist:
- Minute-by-minute compound tracking during real roasting
- Surface vs. core concentration gradients for acids and flavor compounds
- Airflow effects on volatile removal
- Humidity optimization for acid volatilization
Bottom line: Academic research confirms that Chocolate Alchemy's framework is chemically sound and represents best practices for small-batch roasting. The 254-262F EOR recommendation sits in the sweet spot where chocolate development is complete, acids are substantially removed, and off-flavors haven't yet formed.