Beer fermentation stages: what happens inside the tank

Most descriptions of brewing stop at the kettle. The wort is chilled, yeast is added, and some number of days later there is beer. That compression is understandable for casual purposes, but it leaves out everything that actually determines whether the beer is clean, balanced, and consistent lot after lot. Fermentation is not a single event — it is a progression of four distinct phases, each with its own yeast behavior, its own risk profile, and its own window for brewer intervention.

Getting those phases right is not about art or instinct. It is about understanding what the yeast is doing at each moment, what it needs from the brewer, and what can go wrong if the conditions are wrong. A tank of wort pitched with healthy yeast at the right rate, held at the right temperature, and managed through each phase will produce clean, consistent beer. Skip a step or get the conditions wrong, and the off-flavors that result are entirely predictable — and entirely avoidable.

Lag phase: first hours after pitching

After pitching yeast into cooled wort, nothing appears to happen for a while. The airlock is still. The wort looks unchanged. This quiet period — typically 6 to 18 hours — is called the lag phase, and describing it as "nothing happening" is precisely wrong. The yeast is working hard; it is just not yet producing alcohol or CO2 in visible quantities.

What the yeast is actually doing is preparing its cellular machinery for the fermentation ahead. It is absorbing dissolved oxygen from the wort and using it to synthesize sterols and unsaturated fatty acids — the structural components it needs to build membranes that can withstand the osmotic pressure of a fermenting environment. It is scavenging zinc and other trace minerals from the wort. It is building up the enzyme systems — the maltases, maltotrioses, and hexokinases — that it will need to convert sugars to alcohol at scale. None of that shows up in the airlock.

The length of the lag phase is a direct readout of yeast health. A well-pitched batch with fresh, properly propagated yeast at an adequate pitching rate will typically show visible krausen and CO2 activity within 12 to 16 hours. A lag that stretches past 24 hours is a diagnostic signal: the yeast pitched was either under-healthy (high dead-cell count, poor viability after cold storage), the pitching rate was too low for the wort gravity, or the wort was nutrient-deficient — low in zinc, low in free amino nitrogen, or both. An extended lag matters not just because fermentation starts late. It matters because the tank is full of sweet, non-acidic wort at fermentation temperature with no competitive pressure from active yeast. That is an open invitation for lactic acid bacteria and other spoilage organisms to establish a foothold before the pH drops and CO2 builds to inhibitory levels.

Exponential growth phase: peak activity

Once the lag is complete, fermentation accelerates rapidly. CO2 production becomes unmistakable — bubbling through the airlock or blow-off tube, a rising head of krausen foam across the surface of the wort, and a measurable drop in specific gravity that a brewer can track with a refractometer or hydrometer. This is the exponential growth phase, and for a standard ale at fermentation temperature it typically runs from roughly 12 to 72 hours after pitching. For a lager at 7 to 9°C, the same energy plays out more slowly, over 3 to 5 days.

During this phase, yeast cell counts double every few hours. Alcohol and CO2 are produced in roughly equimolar quantities from glucose as the yeast runs glycolysis at full speed. The reaction is exothermic, and in a large tank this matters practically: a 1,000-liter batch of ale at high-krausen can generate enough metabolic heat to raise the beer temperature by 2 to 4°C if the cooling jackets are not actively pulling heat out. That temperature creep is not benign — it accelerates yeast metabolism in ways that produce more stress byproducts, and it is one of the most common causes of fusel alcohol elevation in commercial fermentation.

Most of the compounds that become off-flavors in the finished beer are generated in this phase. Acetaldehyde — which tastes of green apple and raw cider — is an intermediate in the alcohol synthesis pathway; the yeast produces it before converting it to ethanol, and during peak growth, production temporarily outpaces reduction. Diacetyl (the compound responsible for the butter and butterscotch flavor associated with under-conditioned lager) is not produced directly — instead, the yeast excretes alpha-acetolactate, a precursor that oxidizes to diacetyl outside the cell. Fusel alcohols — the higher alcohols that give fermentation-stressed beer a hot, solvent character — arise from yeast amino acid catabolism under conditions of high growth rate, high temperature, or inadequate nitrogen in the wort. All of these are normal metabolic outputs of a yeast working hard; the question is whether subsequent phases clean them up.

Attenuation phase: cleaning up

As the simple fermentable sugars — glucose, fructose, sucrose, and most of the maltose — are depleted, yeast growth rate slows and the fermentation enters a quieter, longer phase. The krausen subsides. CO2 production drops to a slow trickle. Gravity continues to fall, but more gradually. This is the attenuation phase, named for the progressive reduction in wort density as the remaining fermentable sugars are consumed. The yeast's metabolic priority shifts from rapid growth to maintenance, stress tolerance, and — critically — the cleanup of the byproducts it generated during peak activity.

The most important cleanup reaction from a quality standpoint involves diacetyl. The alpha-acetolactate that the yeast excreted during exponential growth has by now oxidized to diacetyl. The yeast reabsorbs diacetyl through the cell membrane, reduces it to acetoin, and then to 2,3-butanediol — a compound with a flavor threshold so high it is effectively flavor-neutral in beer at normal concentrations. This reductive pathway is the sole mechanism by which diacetyl is removed from beer; filtration, finings, and cold conditioning do not reduce diacetyl. Only active yeast in the tank can do it, and only if the yeast is healthy and the temperature is adequate for the enzymes to function.

For lager fermentation, the diacetyl rest is a deliberate brewing step that exploits this biology. After primary fermentation at 7 to 9°C, the brewer raises the tank temperature to 12 to 16°C and holds it there for 24 to 72 hours. The warmer temperature accelerates the enzymatic reduction of diacetyl by the yeast still in suspension, driving it below sensory threshold before the beer is chilled for conditioning. Skipping or shortening this rest — a common pressure point in high-volume commercial production — is the single most frequent cause of the buttery flavor that distinguishes a poorly conditioned lager from a clean one. It is not a subtle defect; most drinkers can detect diacetyl in lager at concentrations as low as 0.1 mg/L.

Conditioning and cold crashing

Once primary fermentation is complete — defined operationally as gravity stable over two consecutive readings 24 to 48 hours apart, at or near the predicted terminal gravity — the beer enters conditioning. The exact form conditioning takes depends on the beer style and the brewery's process, but the goals are consistent: allow remaining yeast activity to complete any residual cleanup, allow the flavor profile to round and integrate, allow the beer to clarify, and bring the dissolved CO2 to the target carbonation level.

For ales, conditioning may be as simple as holding the beer at fermentation temperature in the same tank for a few additional days. The yeast is still present in low but significant numbers, flavor compounds continue to evolve, and any residual acetaldehyde or diacetyl is reduced further. Some ale styles benefit from a brief period at slightly elevated temperature during this phase for the same enzymatic reasons as the lager diacetyl rest.

For lagers, conditioning — the "lagering" that gives the style its name — is a much longer process at near-freezing temperatures, typically 0 to 2°C, sustained for weeks to months. At these temperatures, yeast activity drops to near zero, which means any serious diacetyl reduction must have been completed at higher temperature before chilling begins. What cold conditioning accomplishes is different: protein-polyphenol complexes that cause chill haze precipitate and settle, improving clarity and long-term stability; dissolved CO2 equilibrates to the correct level for the storage pressure; and the flavor profile undergoes a gradual integration and rounding that is difficult to replicate by any faster method. A lager rushed out of lagering before this integration is complete will taste raw, sharp, and one-dimensional compared to the same beer given its full conditioning time.

Cold crashing is a related but distinct step: a rapid drop to near-zero temperature, typically over 12 to 24 hours, that causes yeast to flocculate and drop out of suspension quickly. Most commercial breweries cold crash before filtration or fining to reduce the yeast load the downstream process has to handle. Highly flocculent yeast strains drop out more completely with less time; low-flocculation strains may need fining agents such as isinglass or bentonite to achieve the same clarity. The choice of yeast strain, fining protocol, and filtration method are all downstream consequences of decisions made at the conditioning stage.

What goes wrong and when

Off-flavors are not random. They are mechanistically tied to specific phases of fermentation, and knowing which phase generated a flaw points directly to the corrective action. Acetaldehyde — the green apple, raw cider character that signals an incomplete fermentation — almost always means the yeast was removed from the beer before cleanup was finished. This can happen through premature cold crashing, overly aggressive filtration of a still-active fermentation, or a flocculation event that dropped the yeast out of suspension before attenuation was complete. The fix is time and yeast contact, not anything added downstream.

Diacetyl points to the attenuation phase and specifically to the VDK rest. If the diacetyl rest was too short, too cold, or performed with yeast that was already stressed and losing enzymatic capacity, the reductive pathway stalls and diacetyl remains in the finished beer. In lager production, diacetyl at detectable levels is almost always a process failure in the conditioning step, not an ingredient problem. The corrective action is straightforward: extend the diacetyl rest, raise the temperature slightly, and verify yeast health before the next pitch.

Fusel alcohols — the hot, solvent character that makes some commercial lagers rough at high intake — come from the exponential growth phase. High fermentation temperature and high-gravity wort both drive fusel production upward, because both force the yeast to catabolize amino acids more aggressively for nitrogen. The fix is temperature control during peak fermentation, not a downstream process change. Once the fusels are in the beer, they stay there.

DMS — the cooked corn or creamed corn flavor associated with some pale lagers — is primarily a pre-fermentation problem: inadequate boil vigor or excessive wort cooling time allows S-methylmethionine (SMM) in the wort to hydrolyze to DMS before fermentation begins. But slow, cool fermentation with extended lag can exacerbate DMS, because a vigorous fermentation's CO2 production helps volatilize DMS out of the fermenting beer. A sluggish fermentation loses that scrubbing effect. The fix begins at the kettle and the whirlpool, not in the fermentation vessel — but a fast, healthy fermentation is the second line of defense.

Understanding which phase each flaw originates in is not academic. At the scale of a commercial brewery running hundreds of batches per year, systematic off-flavor diagnosis tied to fermentation phase data is what separates a brewery that learns from failures from one that repeats them. Temperature logs, gravity curves, yeast cell counts, and VDK measurements for every batch form the dataset that makes that diagnosis possible.