Winemaking Laboratory
Alcoholic fermentation / Yeast: biology and selection

Chapter III · Alcoholic fermentation

Yeast: biology and selection

How yeast turns sugar into alcohol, how wild and cultured yeasts succeed one another during fermentation, where every gram of sugar goes, and what kind of strain suits your wine.

Fermentation conditions

Illustrative model: three groups — Hanseniaspora (Kloeckera), other non-Saccharomyces yeasts and Saccharomyces cerevisiae. Growth follows the cardinal temperature model (Rosso), limited by alcohol, SO₂ and total capacity; wild starting populations: 10⁵, 10⁴ and 10² cells/mL.

Saccharomyces dominates from—
Peak of wild yeast—cells/mL
Fermentation ends—
Alcohol—% vol

Fate of sugar

Gay-Lussac equation: C₆H₁₂O₆ → 2 C₂H₅OH + 2 CO₂ (51.1% ethanol, 48.9% CO₂ by mass). Glycerol formation uses ≈ 0.98 g of sugar per 1 g of glycerol; "other" — succinic acid, acetic acid, acetaldehyde, higher alcohols, esters. Ethanol density 0.789 g/mL.

Ethanol—% vol
CO₂—g/L
Sugar per 1% alcohol—g/L per % vol
Of theoretical—%

Wine and conditions

The result is a profile of strain traits, not a specific brand. Look in producers' tables for a strain that meets these traits.

Alcohol tolerance needed—% vol
Temperature range—°C
Nitrogen demand—mg N/L
YAN—mg N/L
Recommended type

Intuition

Yeast makes wine. In a few weeks the single-celled fungus Saccharomyces cerevisiae turns the sugar of must into ethanol and carbon dioxide, producing hundreds of aroma compounds along the way. A millilitre of fermenting must holds up to a hundred million cells — a single qvevri holds more of them than there are people on Earth.

But yeast is not alone in the must. Dozens of species of wild yeast live on the grape skins and in the marani, and fermentation, especially spontaneous fermentation, is a competition among them. Who wins depends on temperature, SO₂, sugar and alcohol. On this page we see how microorganisms succeed one another, where every gram of sugar goes, and how to choose a strain for a particular wine.

History

In 1680 Antonie van Leeuwenhoek saw yeast cells in beer through his microscope, although he did not realise they were alive. In 1810 Joseph Louis Gay-Lussac set out the overall equation of fermentation: sugar → ethanol + carbon dioxide. Between 1857 and 1860 Louis Pasteur proved that fermentation is caused by living yeast, and found that without oxygen yeast consumes far more sugar — "life without air".

In 1897 Eduard Buchner showed that juice pressed from yeast cells also ferments sugar — fermentation is carried out by enzymes. He received the Nobel Prize for this in 1907, and biochemistry as a science starts here. In 1929 Herbert Crabtree described the effect that explains why S. cerevisiae ferments sugar even when oxygen is present. In 1996 S. cerevisiae became the first eukaryote whose genome was fully sequenced.

Yeast biochemistry

In glycolysis a glucose molecule is split into two molecules of pyruvic acid, yielding two ATP and two NADH. Without oxygen the cell must regenerate NAD⁺: pyruvic acid is decarboxylated to acetaldehyde (releasing CO₂), and acetaldehyde is reduced to ethanol by NADH. Overall:

$$C_6H_{12}O_6 \rightarrow 2\,C_2H_5OH + 2\,CO_2 \qquad \Delta G^\circ{}' \approx -235\ \text{kJ/mol}$$

Only part of the energy is stored as ATP; the rest is released as heat (≈ 98 kJ/mol) — the subject of the "thermodynamics" page of Chapter III.

The Crabtree effect: when glucose exceeds 1–2 g/L, S. cerevisiae represses the synthesis of its respiratory enzymes and ferments sugar despite the presence of oxygen. Energetically this is inefficient (2 ATP instead of more than 30), but evolutionarily it pays off: the yeast grabs the sugar quickly, and ethanol poisons its competitors. Later, once the sugar is gone, S. cerevisiae can also oxidise ethanol.

For redox balance, growing cells produce extra NADH, which is consumed by synthesising glycerol. Glycerol is therefore the main by-product of fermentation (5–10 g/L), adding softness and body to the wine.

Growth phases

In a closed system (must, qvevri) a yeast population passes through four phases:

The lag phase (a few hours to a day) — the cells adapt to the new environment and hardly grow. The exponential phase — the population grows at a constant rate: $N(t) = N_0\,e^{\mu t}$, doubling time $t_d = \ln 2 / \mu$, 2–3 hours at the optimum temperature. The stationary phase — growth stops as nitrogen or other nutrients run out and alcohol accumulates, but the cells keep fermenting sugar — most of the fermentation happens in this phase. The death phase — at high alcohol and after the sugar is exhausted the cells die and autolysis begins.

The calculator describes how growth rate depends on temperature with Rosso's cardinal model — three parameters: minimum, optimum and maximum temperature:

$$\mu(T) = \mu_{\text{opt}}\,\frac{(T - T_{\max})(T - T_{\min})^2}{(T_{\text{opt}} - T_{\min})\left[(T_{\text{opt}} - T_{\min})(T - T_{\text{opt}}) - (T_{\text{opt}} - T_{\max})(T_{\text{opt}} + T_{\min} - 2T)\right]}$$

For S. cerevisiae $T_{\min} \approx 4$, $T_{\text{opt}} \approx 32$, $T_{\max} \approx 45$ °C; for wild non-Saccharomyces yeasts the optimum is lower (25–26 °C), and at low temperature they grow relatively better.

Microbial succession

On healthy grape skins S. cerevisiae is very rare — tens of cells per millilitre of must. Hanseniaspora uvarum (formerly Kloeckera apiculata) dominates at 10⁴–10⁵ cells/mL, along with Candida, Metschnikowia, Pichia and Torulaspora. In the first days of a spontaneous fermentation it is they that multiply and begin fermenting the must.

Non-Saccharomyces yeasts tolerate alcohol poorly: they die at 4–6%. S. cerevisiae tolerates up to 14–16%. So in the middle of fermentation, as alcohol rises, S. cerevisiae takes over and finishes the job. SO₂ speeds up this succession: non-Saccharomyces yeasts are suppressed even at 0.2–0.5 mg/L of molecular SO₂, while S. cerevisiae withstands far more.

Non-Saccharomyces yeasts are not simply "harmful": their enzymes release aroma precursors and add complexity to the wine. But some of them produce acetic acid and ethyl acetate (a solvent smell), especially if fermentation starts slowly. That is why modern practice sometimes inoculates selected non-Saccharomyces strains (Torulaspora delbrueckii, Metschnikowia pulcherrima) together with, or before, S. cerevisiae.

Mass balance

By the Gay-Lussac equation, 180 g of glucose gives 92 g of ethanol (51.1%) and 88 g of CO₂ (48.9%). In reality part of the sugar goes elsewhere:

$$S = S_{\text{ferm}} + S_{\text{glyc}} + S_{\text{biom}} + S_{\text{other}}, \qquad m_{\text{eth}} = 0.511\,S_{\text{ferm}}\,(1 - \lambda)$$

where $\lambda$ is the share of ethanol evaporated with the CO₂. Forming 1 g of glycerol uses ≈ 0.98 g of sugar, biomass takes 1–2% of the sugar, and succinic acid, acetic acid, acetaldehyde and other products about 1%. As a result the real yield is 90–95% of theoretical, and 16.5–17.5 g/L of sugar is needed per 1% of alcohol. EU legislation uses 16.83 g/L to compute potential alcohol (see "Sugars and must density").

Converting to alcohol by volume: $\text{alc.}\,(\%\ \text{vol}) = m_{\text{eth}} / 7.894$, where 7.894 g/L is the mass of ethanol per 1% by volume (density 0.789 g/mL).

Secondary products of fermentation

Glycerol (5–10 g/L) — sweetish taste, viscosity, "body"; more forms at high temperature and high sugar. Acetaldehyde (10–100 mg/L) — the precursor of ethanol; in excess an oxidised, "sherry-like" note; binds SO₂. Higher alcohols (150–400 mg/L) — isoamyl, isobutyl and propyl alcohols and 2-phenylethanol (rose aroma); they form from amino-acid metabolism and in moderation add complexity, in excess a coarse, heavy note. Esters — ethyl acetate, isoamyl acetate (banana), ethyl hexanoate (apple); more of them stay in the wine at low temperature. Hydrogen sulfide (H₂S) — "rotten egg"; forms under nitrogen shortage and stress. Acetic acid (0.2–0.5 g/L is normal) — more under the action of wild yeast and bacteria.

Strain selection

Hundreds of selected strains of S. cerevisiae exist, and they differ in several traits: alcohol tolerance (12–18%), temperature range (some ferment even at 10–12 °C), nitrogen demand, production of glycerol, higher alcohols and esters, release of thiols (Sauvignon-type aromas), adsorption of anthocyanins on the cell wall, production of SO₂ and H₂S, and the killer factor — a toxin with which a strain kills sensitive competitors.

The logic of selection follows from the wine style and conditions. Potential alcohol sets the tolerance needed (at least 1% higher), fermentation temperature the temperature range, and must nitrogen the nitrogen demand. If malolactic fermentation is planned, the strain should produce little SO₂, because SO₂ suppresses lactic acid bacteria.

Key quantities

QuantityTypical valueNote
S. cerevisiae on grape skins10 – 10³ cells/mLon healthy grapes
Hanseniaspora on grape skins10⁴ – 10⁵ cells/mLdominates at the start
Maximum population10⁸ cells/mLdepends on nitrogen
Doubling time2 – 3 hat the optimum temperature
Non-Saccharomyces alcohol limit4 – 6%they die
S. cerevisiae alcohol tolerance14 – 16% (up to 18% by strain)—
Theoretical yield0.511 g ethanol / g sugarGay-Lussac
Real yield90 – 95% of theoretical—
Sugar per 1% alcohol16.5 – 17.5 g/Llegal 16.83
Glycerol5 – 10 g/L—
Higher alcohols150 – 400 mg/L—

Connections

"Preparing the must" in the previous chapter is the prerequisite for this page: sulfiting, settling, yeast inoculation and nitrogen are calculated there. The next page, "Fermentation kinetics — simulator", links the growth described here to sugar consumption and computes the fermentation curve day by day; "thermodynamics" covers the heat of fermentation; "Malolactic fermentation" the bacterial fermentation that follows the alcoholic one. Chapter VII (Sensory science) looks at the secondary products of fermentation from the point of view of aroma.

Application — Georgian practice

For centuries traditional qvevri wine was made by spontaneous fermentation. When fermenting on the pomace, the microflora of the grape skins is spread through the whole mass, and the porous walls of the qvevri and the marani itself carry their "own" microflora from year to year. This gives the wine a character of place, and some researchers regard the role of local microflora as part of terroir.

Features of spontaneous qvevri fermentation: the temperature rises to 28–32 °C, which quickly suppresses non-Saccharomyces yeasts and favours S. cerevisiae; the pomace gives the yeast extra nutrients and surface; and if fermentation starts slowly (a cold autumn, high pH), the risk from acetic bacteria and wild yeast rises. That is why many Georgian winemakers use a pure yeast strain even in qvevri, or a "mother" — a small amount of must fermented in advance from healthy grapes, which gives the main mass an active fermenting culture.

Glossary

TermDefinition
Alcoholic fermentationConversion of sugar to ethanol and CO₂ by yeast
GlycolysisThe biochemical pathway breaking glucose down to pyruvic acid
Crabtree effectFermentation in the presence of oxygen at high sugar
Lag phaseThe yeast's adaptation period before growth starts
Exponential phasePopulation growth at a constant relative rate
Stationary phaseGrowth has stopped, fermentation continues
Death phaseDeath of the cells and autolysis
Non-Saccharomyces yeastsWild yeasts (Hanseniaspora, Candida, Pichia etc.)
Killer factorA toxin with which a strain kills sensitive yeast
GlycerolA by-product of fermentation, a component of wine body
Higher alcoholsAlcohols with longer chains than ethanol (isoamyl etc.)
EstersCompounds of acids and alcohols, fruity aromas
"Mother"Must fermented in advance to start the main mass

Practical use in calculations

A must at 22 Brix (≈ 216 g/L sugar) ferments spontaneously at 20 °C without SO₂. The model shows Hanseniaspora dominating in the first days and reaching 3.5·10⁷ cells/mL; on the third or fourth day (≈ 3.3 days) S. cerevisiae takes the lead, and at 4–5% alcohol the wild yeast dies off quickly. Fermentation finishes in about 16 days at 12.8% alcohol. With 0.5 mg/L of molecular SO₂ the wild-yeast peak falls tenfold (≈ 2.8·10⁶), and with a pure strain inoculated (5 million/mL) S. cerevisiae dominates from the start and fermentation finishes in ≈ 14 days. At 14 °C the same must needs ≈ 29 days; at 28 °C a spontaneous fermentation takes ≈ 10. A 26 Brix must at 20 °C, on the other hand, stops with residual sugar when alcohol reaches 15% — the main risk of high sugar.

Mass balance for the same must (glycerol 7 g/L, biomass 1.5%, other 1%, evaporation 1%): 103 g/L of ethanol, i.e. 13.1% vol, and 100 g/L of CO₂ — about 54 litres of gas per litre of must. 16.5 g/L of sugar was used per 1% of alcohol; the yield is 93% of theoretical.

For an amber wine in qvevri (23 Brix, 26 °C, nitrogen 160 mg N/L, malolactic fermentation planned) the selection mode shows: potential alcohol 13.5%, tolerance needed ≥ 14.5%, temperature range 20–30 °C, nitrogen need ≈ 250 mg N/L — nitrogen is short. Recommended: a robust S. cerevisiae with a wide temperature range and low nitrogen demand that produces little SO₂ — or spontaneous fermentation, if the grapes are healthy and control is close.

For beginners — what this page shows

1. In the "Microbial succession" mode choose the fermentation type — spontaneous or with a pure strain — and enter the temperature, sugar and SO₂. The chart shows how three groups of yeast grow and die and how alcohol rises.

2. Watch the "Saccharomyces dominates from" card: the earlier this happens, the smaller the influence of wild yeast on the wine. Try raising SO₂ or changing the temperature and see how it shifts.

3. The "Fermentation ends" card tells you how many days it takes to ferment the sugar. If it shows residual sugar in red, there is a risk of a stuck fermentation — low temperature or too much sugar.

4. The "Mass balance" mode shows where the sugar of the must goes: to ethanol, CO₂, glycerol and other products. It explains why the alcohol you actually get does not exactly match the one calculated in advance.

5. In the "Strain selection" mode choose the wine style and enter the conditions. You get a profile of desirable strain traits — use these traits to find a strain in producers' tables.

For professionals — limits of the models

1. The succession model lumps species into three groups and is illustrative. In real must the species composition depends on grape condition, vineyard treatment and the cellar microflora. The parameters (μ, E50, SO₂ sensitivity) are typical literature values.

2. Nitrogen is not a separate variable in the model — its effect is included through the maximum population. A detailed model of nitrogen deficiency and diagnosis of stuck fermentations is on the next page, "Fermentation kinetics — simulator".

3. The mass-balance parameters depend on one another: higher temperature gives more glycerol and more evaporation, lower temperature more esters. If you have measured the actual alcohol in your wine, adjust the parameters until the model matches — you get the real "sugar per 1% alcohol" of your cellar.

4. The strain-selection rules are a general guide. The same strain behaves differently in different musts; try a new strain on a small volume first.

5. For spontaneous qvevri fermentation, a "mother" — a small volume of must from healthy grapes fermented 1–2 days in advance (2–5% of the main mass) — shortens the lag phase and lowers the risk from wild microflora while keeping the character of the local microflora.

1. Why does Saccharomyces cerevisiae ferment sugar even when oxygen is present?

Crabtree effect: above 1–2 g/L glucose S. cerevisiae represses its respiratory enzymes and ferments sugar rapidly. Ethanol poisons competitors — this is its evolutionary strategy.

2. At the start of a spontaneous fermentation, which yeast is most abundant in the must?

On grape skins S. cerevisiae is very rare (tens of cells/mL), while Hanseniaspora reaches 10⁴–10⁵. Once alcohol exceeds 4–5%, non-Saccharomyces yeasts die and S. cerevisiae dominates.

3. Why doesn't 1 g of sugar give 0.511 g of ethanol, as the Gay-Lussac equation says?

Real yield is 90–95% of theoretical: 4–5% of the sugar goes to glycerol, 1–2% to biomass, ~1% to succinic acid and other products, and part of the ethanol evaporates with the CO₂.

4. For an amber wine in qvevri, in a warm marani, with low nitrogen — which trait matters most?

In a qvevri the temperature rises to 28–32 °C and fermentation on the pomace is long. A robust strain that tolerates heat and low nitrogen reduces the risk of a stuck fermentation.