Chapter 34 — Coffee, Tea, and Chocolate: Fermentation You Didn't Know About

The Hook

Chef Aroon Sornprasit was on the roof of his restaurant, in the cold light of a March morning in Toronto, picking through a small wooden tray of coffee beans. The tray was Thai — sent from a friend in Chiang Rai, north of his home city — and the beans were raw, green, hard as river pebbles, smelling faintly of grass and faintly of fruit and faintly of hay. They came from the Doi Tung Royal Project, a development program in the mountains of northern Thailand where, over the last forty years, opium-poppy farms had been gradually converted to coffee, macadamia, and tea cultivation.

The beans Aroon was holding had not been roasted yet. That was his job, in a small drum roaster downstairs. But before they had reached him, before they had crossed the Pacific Ocean in a container ship, before they had been packed in burlap, before they had been dried in the sun for two weeks, they had been fermented.

This is the part most people don't know. The bean Aroon was holding was a seed. It had grown inside a fruit — a red cherry, sweet on the outside, the size of a small grape — that had been picked from a tree in Doi Tung. The seed sat in the center of the cherry, surrounded by a sticky pulp called mucilage. When the cherry was harvested, it had to be processed. The pulp had to come off. And the way the pulp came off — the way it was removed — depended on whether the wet mucilage layer fermented in the open air for two days, or in tank water for a day, or in sealed anaerobic vessels for a week. The choice was made by people Aroon had never met. The choice was decisive. By the time the green bean reached his hand, the coffee inside had already been half-built.

Aroon weighed out 200 grams. He carried them downstairs. He warmed his roaster, set the airflow, listened for the first crack. The chemistry of the next twenty minutes — the Maillard reactions of Chapter 8, the caramelization of Chapter 10, the aromatic compounds of Chapter 21 — was about to unfold with the same precision and the same beauty as it always did. But the raw material was already a fermented food.

Maya Okonkwo learned this for the first time when Aroon told her, casually, over a cup of his Doi Tung roast in his kitchen. Maya had spent a year reading about coffee origins, single-origin labels, terroir, "Ethiopian heirloom varietals," the difference between washed and natural processing. She had thought processing meant the drying method. It hadn't occurred to her that "washed coffee" and "natural coffee" were two different fermentations, and that the difference between the two cups in front of her — one bright, citric, almost tea-like; the other heavy, fruit-forward, with notes of blueberry and red wine — came not from different beans but from different microbes.

This chapter is about three of the most-consumed foods and beverages on earth: coffee, tea, and chocolate. All three are products of a long supply chain. All three involve fermentation steps that happen on the farm, before the consumer ever sees them, that most consumers don't know about. All three are, in their consumer-facing form, the end of a microbial story that began when the fruit was harvested.

By the end of this chapter, you will not be able to drink a cup of coffee, eat a piece of dark chocolate, or sip a cup of pu-erh tea without thinking about what was happening, on a farm in Ethiopia or Côte d'Ivoire or Yunnan or Sumatra, six months before the cup reached your hand.

The Everyday Observation

Pour yourself a cup of coffee. Look at it. The cup is dark brown, sometimes almost black. The smell is roasted, with notes that might remind you of caramel, chocolate, fruit, nuts, baked bread, or smoke. The flavor is bitter and acidic, often sweet underneath, sometimes savory. None of those compounds — not the brown color, not the roast aroma, not the bitter compounds, not the chocolate notes — were present in the coffee fruit on the tree.

The cherry on the tree is a small red fruit that tastes faintly sweet, mildly fruity, slightly mucilaginous. The seed inside is green and hard and tastes mostly of nothing — slightly grassy, slightly herbal, but flat. There is no chocolate in it. There is no caramel. There is no roast.

Everything we associate with coffee was built. Some of it was built in the roaster (Chapter 8 and Chapter 21 explained those reactions). But some of it — much of it, depending on whom you ask — was built earlier, on the farm, by microbes working on the cherry pulp around the bean. That earlier step is what this chapter is about.

The same is true for chocolate. Crack open a fresh cacao pod. Inside are roughly 30 to 50 seeds (cocoa beans), each surrounded by a sweet white pulp that tastes a little like a cross between mango, lychee, and yogurt. The seeds themselves, raw and unfermented, taste terrible. They are bitter, astringent, harsh, and bear no resemblance to chocolate. The first published European descriptions of cacao after the conquest describe people biting into raw beans and being repulsed. Chocolate as we know it does not exist in the cacao fruit. It exists only after fermentation, drying, roasting, conching, and tempering. The flavor is engineered, in stages, by a long chain of microbial and chemical transformations. The first stage is fermentation.

And tea. The cup of black tea you might drink at breakfast — the kind from Assam or Sri Lanka or Kenya, dark, malty, robust — is a transformation of green leaves that, freshly plucked, taste grassy and astringent. The transformation has historically been called "fermentation" by tea growers. It is, in fact, mostly enzymatic oxidation, not microbial fermentation, but the word has stuck for centuries and we'll have to disentangle it carefully. Some teas, however — especially the dark teas of Yunnan — really do undergo microbial fermentation, by molds and bacteria, over months or years. The cup of pu-erh you might drink with dim sum is microbially aged, in some cases for decades.

Three foods. Three fermentation stories most consumers have never heard. Let us take them in turn.

The Science

Cacao fermentation: the step that creates chocolate

The cacao tree, Theobroma cacao ("food of the gods" in Linnaean Latin, a name lifted from Mesoamerican reverence for the plant), is native to the Amazon basin and was first cultivated by Indigenous peoples of Mesoamerica — the Olmec, then the Maya, then the Aztec. We will not retell the full history of cacao here; Chapter 20 covered cacao botany and the European appropriation of the bean. What we want to do in this chapter is the fermentation step.

The fruit

A cacao pod is the size of a small football, green or yellow or red or purple depending on variety, with a thick rind. Inside are 30 to 50 seeds (the "beans") arranged in a column down the center, each individually surrounded by a sweet, mucilaginous white pulp. The pulp is made of sugars (about 10 to 15 percent by weight, mostly fructose, glucose, and sucrose), citric acid (which keeps the pulp tart and at low pH), and water, plus various pectins, proteins, and trace compounds.

Here is the key fact: the bean inside is bitter, astringent, and flat-tasting in its raw state. It contains the precursor compounds that will, eventually, become chocolate flavor — but those precursors are locked up. To unlock them, you have to ferment the surrounding pulp, and you have to ferment it in a way that kills the bean's germ (the part that would otherwise sprout) and changes the bean's internal chemistry from the outside in.

The procedure: heap, box, or basket

After harvest, cacao pods are split open and the wet seed mass — beans plus pulp, all together — is placed in a fermentation vessel. The traditional vessels vary by region:

• Heap fermentation (West Africa, parts of Latin America). Wet seed mass is piled on banana leaves on the ground, covered with more banana leaves, and left to ferment for 5 to 7 days. Workers turn the heap every 1 to 2 days to expose more of the mass to air. This is the dominant fermentation method in Côte d'Ivoire and Ghana, which together produce roughly 60 to 70 percent of the world's cocoa.

• Box fermentation (much of Latin America, parts of Indonesia, increasingly common in West Africa). Wet seed mass is placed in wooden boxes, often arranged in cascading tiers so that beans can be transferred to a new box every day or two as the fermentation progresses. Box fermentation gives more controlled exposure to oxygen and more even fermentation. It is the standard for higher-end cacao production.

• Basket fermentation and pulp washing (smaller regional variations). In some traditions, beans are partially de-pulped before fermentation; in others, they are fermented in baskets that allow drainage of the liquid runoff (called "sweatings").

The microbial succession

In all these methods, the fermentation proceeds in a remarkably consistent succession of three microbial phases over 5 to 7 days. This is one of the cleanest examples of microbial succession in food science, and it has been documented in detail by researchers in Côte d'Ivoire, Brazil, Ecuador, Indonesia, and elsewhere.

Phase 1 (first ~24 hours): Yeasts. The pulp is acidic (pH around 3.5), low-oxygen (the pulp is sticky and excludes air), and full of sugar. These conditions favor yeasts, especially Saccharomyces cerevisiae (yes, the same species as bread and beer yeast — Chapter 31), Hanseniaspora opuntiae, Pichia kudriavzevii, and others. They consume the pulp's sugars and produce ethanol (alcohol) plus CO₂. The pulp begins to break down — the pectins that gave it structure are partly digested by yeast pectinases. Liquid drains away (the "sweatings"). The temperature begins to rise as yeast metabolism releases heat.

Phase 2 (days 2–4): Lactic acid bacteria. As ethanol accumulates and oxygen begins to penetrate the breaking-down pulp, yeasts decline and lactic acid bacteria — Lactobacillus plantarum and others, the same family that did the work in Chapter 32 and Chapter 33 — take over. They consume remaining sugars and produce lactic acid. The pH starts to drop further; the temperature continues to rise.

Phase 3 (days 4–7): Acetic acid bacteria. As pulp breaks down further and oxygen flows freely through the heap, acetic acid bacteria (Acetobacter and Gluconobacter species) come in. They oxidize the ethanol that the yeasts produced in Phase 1 into acetic acid (vinegar). This is the same chemistry that turns wine into vinegar — and it produces a great deal of heat. Heap temperatures during the acetic acid phase commonly reach 45 to 50°C (113 to 122°F).

That heat, combined with the acetic acid penetrating the bean, is what does the work the fermentation is for. The acid soaks through the bean's seed coat. Inside the bean, several things happen at once:

• The bean's germ is killed by heat and acid. The bean can no longer sprout.
• The bean's internal cell membranes break down, mixing previously separated compounds.
• Stored bean proteins are partially hydrolyzed by the bean's own proteases, which become active under these conditions, producing free amino acids and small peptides.
• Polyphenols — the bitter, astringent compounds that made the raw bean unpalatable — are oxidized and complexed with proteins, substantially reducing astringency.
• Sugars are partly mobilized.

The bean leaves fermentation transformed at the cellular level — its protein and sugar profile rebuilt, its astringency dramatically reduced, its flavor precursors now in place. The bean is then dried (to about 7 percent moisture, on raised drying tables or patios in the sun) and shipped to roasters.

Why the timing matters

Cacao fermentation is one of the most timing-sensitive steps in any food's production. Under-fermented beans (less than about 4 days, or with a stalled fermentation) come out flat, astringent, and harsh. There is not enough acid penetration; the bean's flavor precursors are not in place. Over-fermented beans (past about 7 days, or in a too-hot heap) can come out vinegary, putrid, with off-flavors that no amount of careful roasting will hide. In between is a window of optimal fermentation that experienced cocoa farmers learn to read by smell, by temperature, by appearance, and by cutting open beans during fermentation to assess the color of the cotyledon (the bean interior) — which transitions from purple to brown as fermentation progresses.

Fine-flavor cacao producers — the small farmers and cooperatives whose beans go to craft chocolate makers — increasingly track fermentation with thermometers and detailed logs, and chocolate makers select lots based on fermentation profile as much as on origin. Bean-to-bar craft chocolate, a movement that has grown substantially since around 2010, has put more attention on the fermentation step than the industry had given it for the previous 200 years. Companies like Dandelion (San Francisco), Soma (Toronto, where Aroon sometimes buys), Mast (Brooklyn), and many others publish details of the bean fermentation when they describe a single-origin chocolate.

Industrial chocolate, by contrast, blends beans from many farms and many fermentation conditions, masking the variability with consistent roasting and conching. There is nothing wrong with industrial chocolate at its best — it is what most of the world eats — but the variability of fermentation is, in a sense, engineered out of the consumer experience.

🌍 Cultural Note — On the producers. The cacao economy is one of the most stark examples of a global food chain in which the people who grow the food and the people who eat it are entirely separate. About 60 to 70 percent of the world's cocoa is produced in West Africa, primarily Côte d'Ivoire and Ghana, on small farms (often 2 to 5 hectares). The farmers who do the harvesting and fermentation usually never taste finished chocolate; the consumers who eat finished chocolate usually never meet the farmers. Cocoa farmers in West Africa earn, on average, less than $2 per day. The industry has long-documented problems with child labor, particularly in Côte d'Ivoire (we noted this in Chapter 20). Fair Trade and direct trade certifications attempt to address some of these inequities; bean-to-bar craft chocolate often pays substantial premiums to specific cooperatives. The science of fermentation is one thing; the economics and ethics of who is paid for it is another, and we cannot honestly write this chapter without naming that.

Coffee fermentation: the step nearly all coffee goes through, even if it is not always called fermentation

Coffee comes from the seeds of the Coffea plant, primarily Coffea arabica (about 60 to 70 percent of world production) and Coffea canephora (often called robusta, about 30 to 40 percent). Coffee was first cultivated in Ethiopia, in the highlands of Kaffa and Sidamo and the broader region, by the ancestors of today's Galla (Oromo) people, who continue to cultivate coffee in landrace varieties unique to the region. From Ethiopia, coffee spread to Yemen, where it was first commercialized as a beverage, and then through the Arab world to Europe and onward.

📜 Naming origins specifically. The phrase "coffee was discovered in Ethiopia" is a Western shorthand that erases the people who actually domesticated it. It is more accurate to say: coffee was first domesticated and used by the ancestors of today's Oromo and other peoples of southwestern Ethiopia, in a landscape that still grows coffee in indigenous varieties on small farms. We name them because they are the people whose work is in every cup.

The fruit

A coffee fruit (called a cherry in industry, even though it is botanically a drupe — a fleshy fruit with a stone in the middle) is roughly the size and shape of a small grape, red or yellow when ripe. Inside is a mucilage layer (sweet, sticky, similar to the cacao pulp), then a parchment (a thin papery skin), then the green coffee bean (usually two beans per cherry, flat-sided, lying back-to-back). The bean itself is a hard, dense seed. Like raw cacao, raw green coffee tastes mostly of nothing — grassy, slightly herbal, very flat.

For the coffee bean to become coffee, the cherry's outer flesh and mucilage have to come off. How they come off — and what role microbes play in that removal — is what determines processing method.

Washed (wet) processing

Most "specialty" coffees you encounter are washed processed. The procedure: ripe cherries are pulped (the outer skin and most of the flesh are mechanically removed), then the still-mucilaginous beans are placed in tanks of water for 12 to 48 hours of fermentation. During this fermentation, microbes (yeasts, lactic acid bacteria, acetic acid bacteria, and others — the community varies with region and farm) digest the remaining mucilage, breaking down the pectins that hold it to the bean. After fermentation, the beans are washed (the loosened mucilage rinses away), then dried in the sun on patios or raised beds for 1 to 2 weeks, until the beans reach about 11 percent moisture.

The flavor profile of washed coffees is generally clean, bright, and acidic. Because the mucilage is removed before drying, the beans dry without prolonged contact with fermenting fruit pulp, and the resulting coffee tastes more "clean" and origin-clear — the underlying terroir of the farm is more directly expressed.

Natural (dry) processing

The other major method is natural processing (also called dry processing), the older of the two and the original Ethiopian method. Whole cherries — fruit and all — are spread on patios or raised beds and dried in the sun for 2 to 4 weeks. During this drying period, the entire fruit is fermenting around the bean, with microbial activity, enzymatic activity from the fruit, and slow drying happening simultaneously. Once the fruit is fully dry and the bean inside is at about 11 percent moisture, the dried fruit is mechanically hulled away.

The flavor profile of natural-processed coffees is generally fruit-forward, sweet, and intense. Because the bean has been in extended contact with the fermenting and drying fruit, sugars and aromatic compounds have penetrated into the bean. Naturals often taste of berries (especially blueberry), tropical fruit, and red wine. Many traditional Ethiopian coffees are natural-processed; many specialty Brazilian, Indonesian, and Yemeni coffees are also natural.

The trade-off: natural processing is harder to do well. The bean is in contact with fermenting fruit for much longer, and the risk of off-flavors, mold, and over-fermentation is higher. A great natural coffee is striking; a poor one tastes like fermented compost.

Honey processing (pulped natural)

A middle path, originating in Brazil and now widespread in Costa Rica, Honduras, and elsewhere: honey processing (also called pulped natural). The cherry is pulped to remove the outer skin, but the mucilage is left on the bean during drying. The drying bean is sticky with mucilage (hence "honey"), and microbial fermentation happens during drying rather than in a tank. Result: somewhere between washed and natural — sweet, clean, with some fruit character but less intensity than full naturals.

Within honey processing, sub-categories have developed based on how much mucilage is left: white honey (most mucilage removed), yellow, red, and black honey (most mucilage retained). The amount of mucilage left correlates with how much fermentation activity happens during drying, and thus with the final flavor.

Anaerobic and experimental fermentation

In the last decade, specialty coffee has developed an entire experimental fermentation movement. The general idea: by sealing coffee beans (with or without their mucilage and fruit) in oxygen-restricted vessels for fermentation, with controlled temperature and sometimes added microbial cultures, growers can produce flavors that conventional washed or natural processing wouldn't reliably produce.

• Anaerobic fermentation: Sealed tanks, low oxygen, often with intentional bacterial dominance, often for 36 to 96 hours.
• Carbonic maceration: A technique borrowed from winemaking, where whole cherries are sealed in a CO₂-rich environment and allowed to ferment with intracellular activity.
• Yeast-inoculated fermentation: Specific yeast strains (sometimes from wine or beer industries) are added to cherry or bean mass to push fermentation in particular flavor directions.

These produce some of the most distinctive — and divisive — coffees on the specialty market. Some critics argue the resulting flavors mask the underlying terroir and become a kind of "fermentation cosplay"; some advocates argue they expand the expressive range of coffee. Both views have merit.

Why coffee fermentation matters for the cup

Even washed processing — the "cleanest" of the methods — involves 12 to 48 hours of microbial activity. All coffee is, to some extent, a fermented food. The fermentation degrades the mucilage, develops some flavor compounds, and conditions the bean for drying. The differences between washed, natural, honey, and anaerobic coffees are differences in how the fermentation was managed. Beans of the same variety from the same farm, processed differently, taste like different coffees.

A side-by-side comparison of the four processing methods

To make the choices concrete, here is what an experienced coffee taster typically describes when they cup the same bean from the same farm processed four different ways.

Washed processing, on a single-origin Ethiopian Yirgacheffe: bright, citric, tea-like, sometimes faintly floral. The dominant flavor descriptors are bergamot, lemon, jasmine, black tea, and a clean almost-watery body. The acidity is high and feels lively on the tongue. Aftertaste is short, clean. This is the profile most associated with the Scandinavian and East-Asian "third wave" specialty coffee aesthetic — the coffee equivalent of a clear sake or an unoaked white wine.

Natural processing, the same Yirgacheffe: fruit-forward, jammy, sometimes wine-like, with a heavier body. Descriptors shift to blueberry, strawberry, red wine, and sometimes "boozy" notes. Acidity remains but is less prominent than in the washed; the heavier body and fruit weight dominate. Aftertaste is longer and sweeter. Naturals often divide tasters: some find them gloriously rich, others find them unfocused or "muddy" if poorly executed.

Honey processing (yellow honey, leaving moderate mucilage), same Yirgacheffe: sweet, balanced, somewhere in between. Notes might include caramel, stone fruit, honeysuckle, with a body heavier than washed but lighter than natural. The processing was developed in Costa Rica and Brazil to give specialty buyers a "best of both worlds" option, and the result tends to please a wide range of palates. White honey approaches washed-coffee character; black honey approaches natural-coffee character.

Anaerobic fermentation, same Yirgacheffe: distinctive, sometimes intense, occasionally extreme. Descriptors at the more typical end include exotic-tropical-fruit (mango, pineapple, lychee), spice (cinnamon, clove), and sometimes overt funk (sour beer, pineapple juice that's been sitting). At the extreme end, anaerobics can taste of buttered popcorn, bubblegum, or whisky — flavors that are genuinely unusual in coffee and that some tasters love and others find overwhelming. The flavor depends heavily on what microbial population dominated in the sealed tank, which depends on inoculation choices and tank conditions; this is why anaerobics vary more from lot to lot than washed or natural coffees from the same farm.

The same exercise can be run with cacao. Bean-to-bar makers increasingly publish the fermentation conditions on their packaging: "Tanzania, Kokoa Kamili cooperative, 6-day box fermentation, 14-day raised-bed sun drying" tells you the bean came from a known farm and was managed through a documented fermentation. Two bars from the same cooperative — one from a 5-day fermentation, one from a 7-day fermentation — will taste meaningfully different, with the longer fermentation typically producing more pronounced fruit and acid notes, the shorter producing a milder, sometimes more cocoa-direct profile. As the craft chocolate movement has matured, makers and producers have begun running deliberate fermentation experiments — varying time, vessel, temperature, even adding particular yeast cultures — in a way that parallels the experimental fermentation movement in specialty coffee. The science is the same; the substrate (cacao versus coffee) and the aesthetic goals differ.

Then those beans are dried, shipped, roasted (Chapter 8 and Chapter 21 covered the Maillard and caramelization chemistry of roasting), ground, and brewed. The fermentation-built compounds interact with the roast-built compounds to produce the final cup. A washed Yirgacheffe and a natural Yirgacheffe come from the same Ethiopian highlands; their flavor differences come almost entirely from fermentation.

A note on decaffeination

Most decaffeination methods do not involve fermentation directly, but they happen at the green-bean stage and can affect flavor. Briefly:

• Swiss water process: beans soaked in hot water, caffeine extracted via activated carbon filtration, water-with-other-flavor-compounds returned to beans. No solvents.
• CO₂ process: supercritical CO₂ selectively extracts caffeine without affecting most flavor compounds.
• Solvent processes (methylene chloride, ethyl acetate): chemical solvents bind caffeine. The "natural" sub-variant uses ethyl acetate from sugarcane; the "Mountain Water Process" uses Mexican spring water plus ethyl acetate.

All of these reduce caffeine to less than 3 percent of original (typically <0.1 percent). All affect flavor to varying degrees, with Swiss water and CO₂ generally considered the gentlest.

Tea: oxidation, sometimes fermentation, mostly the former

Tea comes from the leaves of Camellia sinensis, a plant native to southwestern China, particularly the Yunnan region, and to parts of Assam in northeastern India. Camellia sinensis var. sinensis (the Chinese variety) has small, hardy leaves; Camellia sinensis var. assamica (the Assamese variety) has larger leaves and a more robust profile. The use of tea as a beverage in China is documented for at least 4,000 years. The tea plant was domesticated and developed by people in southwestern China — specifically Yunnan and Sichuan, in regions historically inhabited by various peoples including the Bulang, Dai, Yi, and others. As with coffee and cacao, naming the actual originators matters more than the standard "discovered in China" shorthand.

The basic categories

All tea — green, white, yellow, oolong, black, dark — comes from the same species. The differences come from how the leaves are processed after picking. There are six main categories, traditionally classified by Chinese tea culture:

• Green tea (lǜ chá, 绿茶): un-oxidized, kill-greened by heat (pan-firing or steaming) immediately after picking.
• White tea (bái chá, 白茶): minimally processed, dried slowly with very light oxidation.
• Yellow tea (huáng chá, 黄茶): like green but with a brief covered fermentation step (mostly enzymatic, slight microbial).
• Oolong (wū lóng, 乌龙): partially oxidized, between green and black.
• Black tea (hóng chá, 红茶, "red tea" in Chinese): fully oxidized.
• Dark tea (hēi chá, 黑茶): fully oxidized and microbially fermented, often aged.

"Fermentation" and oxidation: untangling the words

Here is where terminology becomes confusing, because for centuries the tea industry has called the oxidation step "fermentation" — and for most tea, this is technically inaccurate.

For green, white, yellow, oolong, and black teas, the transformation is enzymatic oxidation, not microbial fermentation. When tea leaves are picked and bruised (or rolled, or otherwise damaged), the cells release polyphenol oxidase (PPO) — the same enzyme we met in Chapter 13 (PPO browning) and in fruits and vegetables — which begins to oxidize the leaf's polyphenols, especially catechins. The catechins (mostly EGCG, EGC, ECG, EC in green tea) are oxidized to theaflavins and thearubigins, two classes of larger oxidized polyphenols that are responsible for the dark color, the astringency, and much of the flavor of black tea. The tea also produces a wide range of aromatic compounds during this process — terpenes, aldehydes, esters, alcohols — that build the floral, malty, fruity notes of oxidized tea.

This is enzymatic oxidation. It does not require microbes. It does not produce alcohol or acid. The leaf's own enzymes act on the leaf's own polyphenols. The chemistry is closely related to the browning of a cut apple (Chapter 13), in fact, and tea oxidation is essentially controlled apple-browning at scale, with the apple replaced by tea leaves and the goal replaced by flavor development instead of prevention.

Why has the industry called it "fermentation" for centuries? Because before microbiology existed as a science, any slow biochemical transformation that produced a strong flavor change was called fermentation. The word was a placeholder for "something happens in this leaf pile that makes the leaves change color and taste different over time." The term stuck. Today's scientific tea writers increasingly distinguish oxidation (for green-to-black processing) from fermentation (for dark teas, which we'll address next).

Black tea oxidation, in summary: picked leaves are wilted to reduce moisture, rolled or "CTC processed" (cut-tear-curl) to bruise them and release enzymes, then spread out for oxidation at controlled temperature and humidity for 1 to 4 hours, then "fired" (heated) to denature the enzymes and stop the oxidation, then dried. The oxidation period is what determines the color and flavor: short for oolong, longer for black tea.

Pu-erh and dark teas: the genuinely fermented teas

But here is where actual microbial fermentation does enter the tea picture. The dark teas of China — most famously pu-erh (普洱) from Yunnan, but also Liu Bao, Fu Brick, Tibetan tea, and others — undergo microbial fermentation by molds, bacteria, and yeasts, often over months to decades.

There are two main types of pu-erh:

Sheng pu-erh (生普洱, "raw" pu-erh): made by a procedure similar to green tea (pan-fired to deactivate enzymes, rolled, sun-dried), then compressed into cakes, bricks, or other shapes, and aged. As the cake ages, slow microbial fermentation by molds and bacteria — happening at the moisture levels naturally present in the compressed leaf and at room or cellar temperatures — gradually transforms the tea over years to decades. A young sheng pu-erh tastes like a slightly oxidized green tea, somewhat astringent, often vegetal. A 20-year sheng pu-erh tastes deep, woody, almost mushroom-like, with no astringency at all. The transformation is real, slow, and considered analogous to wine aging in Chinese tea culture.

Shou pu-erh (熟普洱, "ripe" pu-erh): a relatively modern (1973) innovation that simulates the long aging of sheng in about 45 days. Tea leaves are piled in heaps and intentionally inoculated with humidity and warmth, allowing rapid microbial fermentation primarily by Aspergillus niger and other molds, plus bacteria and yeasts. The pile reaches temperatures of 50–60°C, and the leaves are turned periodically. After 45 to 60 days of pile fermentation, the result is a tea that mimics, but does not fully replicate, the depth of long-aged sheng. Shou pu-erh is darker, smoother, and more affordable than aged sheng; it is what most people drink with dim sum.

The microbial communities of pu-erh have been studied in some detail, with Aspergillus niger, A. luchuensis, Penicillium, Blastobotrys, and various bacterial species (including Bacillus subtilis and several lactic acid bacteria) consistently identified across cakes from different Yunnan tea factories. The molds break down the leaf's catechins and produce a range of new compounds, including statins (which has led to tentative medical interest in pu-erh's metabolic effects), theabrownins (a class of dark-pigmented polymers different from theaflavins/thearubigins), and various flavor-active esters and aldehydes.

Aged sheng pu-erh, in particular, can appreciate substantially in monetary value over decades — a tea from a famous Yunnan tea mountain in 1990 might sell for tens of thousands of dollars per cake today. The tea collector culture in China and Southeast Asia is, in some ways, parallel to wine collecting, with the same questions about authentication, terroir, and the difference between aged value and aged quality.

The sheng-versus-shou distinction in more practical detail

For the curious tea drinker, the difference between sheng and shou pu-erh is worth understanding in a little more practical detail, because it determines how you buy the tea, how you store it, and how you brew it.

Sheng pu-erh ("raw," 生) is, in essence, a pressed green tea that ages naturally over years to decades through slow microbial action at the moisture levels naturally present in the compressed leaf. A young sheng (under 5 years) tastes recognizably green-tea-like — vegetal, bitter, sometimes astringent, with a sharpness that can feel like the tea is "biting" the back of the tongue. As the cake ages, the bitter compounds (catechins, especially EGCG) slowly transform into mellower oxidized derivatives; the astringency diminishes; deep, woody, almost forest-floor or aged-leather notes emerge. A 30-year sheng from a respected Yunnan mountain (Banzhang, Yiwu, Bingdao, and other named regions) tastes nothing like its young self — it is dark, smooth, often with a sweet aftertaste called huí gān (回甘, "returning sweetness") that lingers in the throat for many minutes after swallowing. The transformation is real and slow and, in tea-collector cultures, considered analogous to the maturation of fine wine. Sheng is bought young (more affordable) and aged at home in a pumidor (a humidity-controlled cabinet kept at 60–70% RH and 20–25°C, named by analogy with cigar humidors), or bought already-aged from specialist vendors at a substantial price premium.

Shou pu-erh ("ripe," 熟) is the modern shortcut: rather than waiting decades for a sheng to mellow, the wò duī (渥堆, "wet pile") technique developed at the Kunming Tea Factory in 1973 simulates much of the aging process in about 45 days. Tea leaves are piled in a damp room, kept at high humidity, occasionally turned, and allowed to ferment under the action of a managed microbial community dominated by Aspergillus niger and related molds. The pile temperature rises to 50–60°C, the leaves darken substantially, and the tannins are aggressively transformed. After 45 to 60 days, the leaves are sun-dried, sorted, and pressed into cakes. The resulting tea tastes deep, smooth, earthy, often with notes of damp wood, tobacco, leather, and dark chocolate. Shou is not the same as long-aged sheng — connoisseurs distinguish them readily — but it is much more affordable and immediately drinkable, and it is the pu-erh most people in the West (and in dim sum restaurants worldwide) actually drink.

A cake of shou pu-erh costs perhaps 20 to 50 USD; a cake of 20-year aged sheng from a named mountain might cost 500 USD or more; a cake of 50-year aged sheng from a famous tea mountain in pristine condition can cost tens of thousands. Authentication is a serious concern in the high-end market — fake "aged" sheng (young tea artificially weathered to look older) is a known problem, and serious collectors rely on trusted vendors and direct relationships with specific Yunnan producers, much as wine collectors rely on auction houses and provenance documentation.

Both sheng and shou are typically brewed gōng fū style — a small clay teapot (often Yixing zisha clay, prized for absorbing tea oils over years and "seasoning" to a particular tea), boiling water, very short infusions (5 to 30 seconds for the first several brews, lengthening as the leaves give up their compounds), and many infusions from the same leaves (10 to 20 is normal for a good pu-erh). The first brew is typically discarded as a "rinse" — it wakes up the compressed leaves and washes off any storage dust. This brewing style maximizes the complexity of what you can extract from a small amount of leaf, and it is a meditative practice in itself across Chinese, Taiwanese, and Korean tea cultures.

Kombucha

A final note belongs to kombucha, which we encountered briefly in Chapter 32. Kombucha is sweetened tea (most commonly black or green) fermented by a symbiotic culture of bacteria and yeasts (the "SCOBY"), where yeasts produce ethanol from the sugar, and acetic acid bacteria oxidize the ethanol to acetic acid, with various other bacteria (sometimes including Komagataeibacter species) producing the characteristic cellulose mat that floats on the brew. Kombucha is genuinely a microbial fermentation of tea — but here the tea is functioning as a nutrient substrate for the SCOBY, more than the microbes are transforming the tea itself. The market for kombucha grew from near-zero in 2000 to several billion dollars annually by the mid-2020s, an example of how a centuries-old home preparation became a commercial category.

🔬 Advanced Sidebar — The microbial succession of cacao fermentation, in detail

For the food-science student or microbiology-curious reader, here is the level of detail at which cacao fermentation has been studied. The dominant recent reference work is Schwan and Wheals (2004) and the substantial follow-up literature from labs in Brazil, Côte d'Ivoire, Belgium, and Indonesia.

Hours 0–24 (pH ~3.5, low O₂, ~25°C): Yeasts dominate. Hanseniaspora opuntiae and Hanseniaspora uvarum in early hours; Saccharomyces cerevisiae and Pichia kudriavzevii increasingly present by hour 24. Yeasts consume pulp sugars (mostly fructose and glucose), produce ethanol (~5 percent of pulp by hour 24) and CO₂. Pectinase activity from yeasts begins breaking down pulp pectin, allowing pulp to liquefy and drain.

Hours 24–72 (pH ~4.0, increasing O₂, 30–40°C): As pulp drains and oxygen penetrates, lactic acid bacteria — especially Lactobacillus plantarum and Lactobacillus fermentum — rise. They consume remaining sugars and produce lactic acid (alongside continued yeast activity, declining). Heap temperature rises from yeast and bacterial metabolism.

Hours 72–168 / days 3–7 (pH 4.0–5.0 rising, full O₂, 45–50°C): Acetic acid bacteria — especially Acetobacter pasteurianus and Gluconobacter oxydans — dominate. They oxidize ethanol from Phase 1 to acetic acid. The reaction is highly exothermic; this is why the heap heats up significantly. Acetic acid penetrates the bean. Bean internal pH drops; bean enzymes (proteases, glycosidases) activate; bean undergoes the internal transformations described above.

Bean-internal chemistry during this period: The bean's storage proteins (especially vicilin-type seed storage proteins) are partially hydrolyzed by bean proteases at slightly acidic pH, producing free amino acids and small peptides (the precursors to Maillard flavor in roasting). Polyphenol oxidase activity oxidizes catechins and procyanidins, reducing astringency. Stored sugars are partly mobilized. By the end of fermentation, the bean's interior color has shifted from purple-violet (raw) to brown (well-fermented), a visible cue cocoa farmers use to assess fermentation progress.

The fermentation flavor precursors then sit, dormant, in the dried bean — until roasting (Chapter 8) drives the Maillard reaction between those amino acids and reducing sugars, building the complex flavor we recognize as chocolate.

🔬 Advanced Sidebar — The chemistry of pu-erh microbial transformation

For the curious tea drinker or microbiology-curious reader: pu-erh microbial fermentation involves a smaller but more selective community than cacao or coffee fermentation, because the substrate (compressed dried tea) is much harsher — low water activity, high tannin, long timescale.

The dominant organism in shou pu-erh pile fermentation is Aspergillus niger (and the closely related A. luchuensis), a black-spored mold that tolerates high tannin levels and produces a wide array of extracellular enzymes including pectinases, cellulases, tannases, and proteases. A. niger is also widely used in industrial citric-acid production and in the koji-fermented foods of southern China and Southeast Asia (compare to A. oryzae in Japanese miso, Chapter 33).

A. niger and its co-fermenters work on the tea's polyphenols and proteins to produce a class of compounds called theabrownins — large dark-colored polymers that are responsible for much of pu-erh's red-brown brewed color and its smooth, less-astringent flavor. They also produce gallic acid (from tannin hydrolysis by tannases) and various small-molecule flavor compounds, including statins (lovastatin in particular has been measured at low levels in some shou pu-erh samples).

The microbial community of aged sheng pu-erh is more complex and slower-evolving — bacteria including Bacillus species are involved, as are various yeasts and lactic acid bacteria, and the community shifts over years. The full ecology has been a subject of active research; the tea industry has only begun, in the last 15 years, to study its own oldest products with modern microbiological tools.

The Practical Application

What does all of this mean for the person standing in a coffee shop or a chocolate aisle or a tea section?

For coffee:

• "Single origin" tells you the farm. The processing method tells you what the microbes did. A washed Ethiopian Yirgacheffe and a natural Ethiopian Yirgacheffe are both "single origin Yirgacheffe" — but they will taste very different because they were fermented differently. When you read a coffee bag, look for "washed," "natural," "honey," or "anaerobic" — and now you know what those words mean.

• Light roasts foreground origin and processing. Dark roasts foreground roast. Chapter 21 explained roast level. The fermentation character of a coffee is most audible in lighter roasts — a light-roast natural will taste of fruit; a light-roast washed will taste of citrus and tea. As roast level deepens, the roast-built flavors increasingly dominate over the fermentation-built ones, until in a very dark roast almost any origin tastes mostly like roast.

• Specialty coffee is paying more attention to fermentation than ever. Producers are publishing fermentation logs. Roasters are buying micro-lots from specific fermentation batches. The "third wave" of coffee culture (espressos beyond Italian-roast tradition, single-origin pour-over, cuppings as serious flavor analysis) increasingly turns on fermentation choices.

For chocolate:

• Bean-to-bar craft chocolate often discloses fermentation details; industrial chocolate rarely does. If a chocolate label says "fermented for 6 days, sun-dried for 14 days, beans from the X cooperative in Y region," you are holding chocolate from a maker who cares about the microbes. If a chocolate label says only "ingredients: cocoa, sugar, ..." you are usually holding industrial chocolate from blended fermentation.

• The bitterness of dark chocolate is partly fermentation-controlled. Well-fermented beans are less astringent than poorly-fermented ones. A craft 70% bar from a well-fermented lot can taste much less bitter than a mass-market 70% bar — not because the cocoa percentage is different, but because the fermentation was better.

• Cocoa nibs you can buy at a health-food store are roasted, fermented cocoa beans, broken up. They are an interesting way to taste the fermentation directly. They are bitter and complex, with no sugar or cocoa butter added.

For tea:

• Most teas you drink are oxidized, not microbially fermented. Black tea, oolong, even most "fermented" teas in Western marketing are oxidized. The chemistry is enzymatic, not microbial. The word "fermented" on a tea box is often shorthand for oxidized.

• Pu-erh is genuinely fermented, and aged sheng appreciates over decades. A pu-erh cake from 1990 is a different drink than the same cake at production. The microbial community has done years of slow work.

• Kombucha is a fermented food made from a tea substrate. The fermentation is real microbial work; the tea is the medium more than the subject.

🍳 Kitchen Lab — The same bean, two ferments (inline tease). This is the most striking comparison-tasting exercise you can do at home. Buy two single-origin coffees from the same farm or region, in the same roast level, with different processing: one washed, one natural. (Most specialty coffee shops will have at least one of each from current rotations.) Brew them the same way (same dose, same grind, same water). Taste them side by side. Note the differences in body, acidity, fruit, and cleanness. The differences are entirely from fermentation. The full comparison-tasting protocol — including a similar exercise for chocolate and pu-erh — is in exercises.md. ⚠️ Allergens: none in pure coffee; cacao products may have dairy (milk chocolate) and soy (lecithin); some cocoa is processed in facilities with tree nuts and peanuts (cross-contamination labeled). Tea: usually none, but some flavored or blended teas contain dairy, soy, or wheat (matcha lattes).

A note on labor, fairness, and where the cup actually starts

The supply chains of coffee, cacao, and tea are some of the most studied — and most fraught — global supply chains. Producers in Ethiopia, Côte d'Ivoire, Yunnan, Sumatra, and elsewhere have, for most of the last 200 years, captured a small fraction of the eventual retail value of their crops. Industries built in the colonial period (cocoa in West Africa, tea in Assam and Sri Lanka, coffee in many Latin American countries) carried inherited inequities into the present.

The last 30 years have seen partial corrections: Fair Trade certification (and, more recently, direct trade and "specialty" labels that pay substantial premiums), the rise of producer cooperatives, and the bean-to-bar / specialty-coffee / single-origin tea movements that, at their best, route more value to growers. None of these have solved the fundamental disparities. A cup of $5 specialty coffee may pay the farmer $0.05 to $0.50; the rest is processing, shipping, roasting, retail, and rent.

This is not a chapter on supply chain ethics. But you cannot honestly write a chapter about cacao fermentation without naming that the work happens in places where people are often badly paid for excellent work, and you cannot honestly write a chapter about coffee fermentation without naming that the experimental anaerobic micro-lots that command $80 per pound at retail in New York and Tokyo are, on the producer end, produced by farmers earning very little of that. The science is the science. The economics are an additional layer of facts.

If this matters to you, look for direct trade sourcing on coffee bags (small roasters often disclose farm-level relationships); look for bean-to-bar craft chocolate with named cooperatives; look for specialty tea vendors who buy directly from family farms in Yunnan or Anhui or Wuyi. These are partial fixes, not solutions. They are also, on average, substantially better deals for producers than industrial alternatives.

🌍 Cultural Note — Naming the people whose work is in the cup.

It is worth restating, plainly, who developed these three foods, because the standard short version ("coffee from Ethiopia, tea from China, cacao from the Americas") flattens out long, specific histories of human work and dilutes credit into geography.

Coffee was domesticated in the highlands of southwestern Ethiopia, in the regions historically inhabited by the Galla / Oromo peoples and other groups of the Kaffa, Sidamo, Yirgacheffe, and Harar regions. The wild Coffea arabica tree still grows in the Ethiopian highland forests; landrace varieties cultivated by Oromo and other Ethiopian farmers are the genetic origin of every arabica bean grown anywhere on earth. The jebena buna (Ethiopian coffee ceremony), in which green beans are roasted, ground, and brewed in a clay pot in front of guests, is one of the oldest continuously-practiced coffee preparations in the world, and the global coffee culture descends, every drop of it, from this Ethiopian root.

Tea was domesticated in the borderlands of present-day southwestern China — particularly Yunnan and Sichuan provinces, in regions historically and currently inhabited by Bulang, Dai, Yi, Hani, Wa, Lahu, and other peoples, alongside Han Chinese populations. The oldest known cultivated tea trees, several of which are over 1,000 years old and still producing leaves, grow on these mountains. The compressed-tea tradition that became pu-erh has Bulang and Dai roots; the broader Chinese tea-drinking culture that spread across East Asia developed across centuries of regional exchange. Indian tea — Assam in particular — was discovered by British colonists in the 1820s, who found wild Camellia sinensis var. assamica growing in tribal lands of the Singpho and other groups in northeastern India, and built the colonial Assam tea industry on a foundation of indigenous knowledge and forced labor.

Cacao was developed by the Indigenous peoples of Mesoamerica — the Olmec, Maya, Mexica/Aztec, and many other groups — over at least three millennia in what is now southern Mexico and Central America, with archaeological evidence of cacao processing extending into Ecuador and Honduras. Every step of cacao processing we have described in this chapter — the pile fermentation, the sun drying, the roasting, the grinding into a paste — is Mesoamerican Indigenous technology, transmitted across generations of farmer-experimentation, and continues to be practiced today in Mesoamerican cacao-growing communities.

The point of naming these people specifically is not symbolic. It is true. The microbes that ferment the bean did not write a manual; the manual is in the practical knowledge of the farmers who learned, by doing it for a hundred or a thousand years, how the heap should smell and feel. That practical knowledge is the original food science.

Cross-Chapter Connections

🔗 Chapter 8 (The Maillard Reaction) is the chemistry of roasting that turns fermented coffee beans, fermented cocoa beans, and (less so) tea leaves into the brown, aromatic foods we recognize. The fermentation builds the precursors; the roasting builds the final compounds.

🔗 Chapter 13 (Enzymes) introduced polyphenol oxidase. Tea oxidation is essentially controlled PPO browning at scale, with the apple replaced by tea leaves and the goal redirected to flavor instead of prevention.

🔗 Chapter 20 (The Science of Chocolate) treated cacao botany, bean structure, conching, cocoa butter polymorphism, and tempering. We deferred the fermentation step there — and now this chapter has it. With Chapter 20 plus this one, you have the full bean-to-bar arc.

🔗 Chapter 21 (The Science of Beverages) treated coffee brewing extraction, tea brewing parameters, and beverage chemistry generally. This chapter explains where the fermented raw material that Chapter 21's brewing is acting on came from.

🔗 Chapter 30 (What Is Fermentation?) named the three categories — yeasts, bacteria, molds — and the framework — selection, succession, endpoint — that this chapter applies to coffee, cacao, and tea.

🔗 Chapter 32 (Cheese, Yogurt, Cultured Foods) introduced the lactic acid bacteria and Lactobacillus plantarum that show up again in coffee fermentation Phase 2 and cacao fermentation Phase 2. The same family of organisms is doing similar work on different substrates.

Looking forward:

🔗 Chapter 38 (The Future Kitchen) will return to coffee and cacao in the form of precision fermentation alternatives — companies attempting to produce coffee-flavored compounds and cocoa-equivalent fats in fermenter tanks, without the agricultural supply chain. Whether these alternatives can reproduce what 5,000 years of farmer-microbe-bean co-evolution has built remains to be seen. Chapter 38 will treat the question seriously.

Closing Reflection

Walk into any coffee shop tomorrow. Order a cup of whatever the barista recommends. Sit down with it.

The cup in your hand is the end of a chain: a tree on a hillside in Ethiopia or Colombia or Indonesia, a cherry harvested by a farmer whose name you will never know, a bin of cherries pulped or piled or sun-dried, a tank of fermenting beans where yeasts and lactic acid bacteria and acetic acid bacteria did 12 to 96 hours of work, a patio of beans drying in the sun, a shipping container, a roaster, a bag, a grinder, a brew. Half of what you taste was built by microbes. The other half was built by heat. The cup is a layered project, and you are tasting all the layers at once.

Aroon's first cup of his Doi Tung roast, the morning he picked through the green beans on the roof, he drank standing at the kitchen window. He set down the empty cup and said, almost to himself: fifteen years ago this was a poppy field. Then, because Aroon is not given to long speeches: now it is breakfast.

The lineage of every food in this chapter — coffee from the Oromo, cacao from the Olmec and Maya, tea from the peoples of Yunnan — runs deep into the human past. The microbes have been doing the work since long before we knew what microbes were. The traditions have been passed through the hands of farmers who taught their children how the heap should smell on day three, how the cherry should give under pressure when it's ready, how the tea pile should feel at the bottom on day twenty. The science we have laid out in this chapter is the science those traditions already knew.

This is the last chapter of Part V. We have come a long way from the first jar of cabbage on the counter, through bread and beer, cheese and yogurt, kimchi and miso, coffee and cacao and pu-erh. The thread is the same: under the right conditions — and humans, across cultures, across millennia, figured out the right conditions — invisible organisms can make food that humans alone cannot. Fermentation is biotechnology that does not need a textbook. The textbook is in the food.

In Part VI, we zoom out. The next chapters are about food safety (the danger of fermentation and cooking gone wrong), food preservation (the larger family of techniques that fermentation belongs to), nutrition science (honestly), and the future of the food system. The microbes will not disappear — Chapter 38 returns to them — but the lens widens. Cooking is not just chemistry. It is also stewardship, public health, ecology, and ethics. Turn the page.