Chapter II · Grape chemistry and harvest
How much sulfur dioxide a must needs and why the dose depends on pH, how long the must takes to clear in a tank, and how much yeast and nitrogen the fermentation will need.
Molecular SO₂ = free / (1 + 10^(pH − 1.81)). The bound fraction in must is usually 30–50%, more in rotten grapes. Wick rate — about 2.5 g of sulfur per 100 L of empty vessel (a practical guide). EU limit for total SO₂ in finished dry wine: red 150, white/rosé 200 mg/L.
Stokes' law: v = 2r²(ρₚ − ρ)g / 9μ. Must viscosity is computed from sugar and temperature; pectin raises viscosity about 1.4-fold and stabilises particles colloidally — without enzyme the model uses an effective diameter 0.6 times smaller. Onset of wild fermentation: ≈ 24 h at 20 °C, Q₁₀ = 2 (a guide).
Nitrogen
Rehydration: water 10 times the yeast mass, 35–40 °C, 15–20 min; when adding to the must the temperature difference should not exceed 10 °C. Nitrogen need — a common guide: ≈ 200 mg N/L at 21 Brix, +25 for each extra Brix. DAP (diammonium phosphate) contains 21.2% nitrogen; EU limit — 1 g/L.
Freshly pressed must is a living system. It contains dozens of species of yeast and bacteria carried in from the grape skins, the vineyard soil and the cellar equipment; enzymes that brown the must on contact with air; and particles — fragments of skin, flesh and soil. Before fermentation begins, the winemaker decides to whom to entrust the must: the wild microflora or a selected pure yeast strain, and under what conditions.
Preparation involves three steps. Sulfiting suppresses unwanted microorganisms and protects the must from oxidation. Must settling frees the must from coarse particles — important for the cleanliness and aroma of white wine. Yeast inoculation and nitrogen supply drive the fermentation quickly and to completion.
One thing is important to understand: sulfiting does not sterilise the must. A normal dose cannot kill every yeast — it selects. Wild yeasts and bacteria are sensitive to SO₂ and are strongly suppressed, while Saccharomyces cerevisiae is far more tolerant. So a pure strain inoculated in sufficient numbers after sulfiting quickly takes over the environment and outcompetes the wild microflora.
Burning sulfur to clean wine vessels has been known since antiquity. In medieval Europe the practice spread widely, and in 1487 a Prussian royal decree officially permitted burning sulfur in wine casks. Potassium metabisulfite powder and sulfurous acid solutions became established in the 20th century — they allow precise dosing.
In the 1860s Louis Pasteur proved that fermentation is caused by living yeast cells. In 1883 the Danish mycologist Emil Christian Hansen isolated the first pure culture of brewer's yeast at the Carlsberg Laboratory, and a few years later Hermann Müller-Thurgau introduced pure yeast strains into winemaking. Active dry wine yeast spread from the 1960s, and today hundreds of selected strains exist.
The physics of particle settling was described in 1851 by George Stokes, who derived the law of drag on a sphere moving through a viscous liquid. This law still underlies every calculation of clarification for wine, beer and water.
In water SO₂ forms three species whose shares depend on pH: molecular SO₂ (dissolved gas), the bisulfite ion $HSO_3^-$ and the sulfite ion $SO_3^{2-}$. The equilibrium between them is set by the constants $pK_{a1} \approx 1.81$ and $pK_{a2} \approx 7.2$. At wine and must pH (3.0–3.8) almost all free SO₂ is in the bisulfite form, and the molecular form is only a few percent:
$$SO_2^{\text{mol}} = \frac{SO_2^{\text{free}}}{1 + 10^{\,pH - 1.81}}$$
It is the molecular form that is antimicrobial: it passes through the cell membrane and disables enzymes inside. At pH 3.0 the molecular share is ≈ 6%, at pH 3.4 ≈ 2.5%, and at pH 3.8 about 1%. In other words, a high-pH must needs several times more SO₂ for the same protection — one of the main reasons why controlling pH (see "Acids, pH and the buffer system") is a precondition for sulfiting.
Part of the added SO₂ quickly binds to carbonyl compounds in the must — acetaldehyde, pyruvic and α-ketoglutaric acids, and partly glucose — and loses its antimicrobial action. This part is called bound SO₂, the rest free. Their sum is total SO₂, and that is what legislation limits:
$$SO_2^{\text{total}} = SO_2^{\text{free}} + SO_2^{\text{bound}}$$
In healthy must 30–50% of the addition binds; in rotten grapes more, because botrytis and bacteria accumulate carbonyls. The calculator takes the bound fraction as a parameter and computes the dose that gives the desired molecular SO₂:
$$D = \frac{SO_2^{\text{mol}}\,(1 + 10^{\,pH-1.81})}{1 - b}$$
where $b$ is the bound fraction.
The second action of SO₂ is antioxidant. It inhibits the enzyme polyphenol oxidase (tyrosinase), which starts oxidising phenolics and browning the must as soon as the grapes are crushed. In rotten grapes the botrytis enzyme laccase acts, and it is far more resistant to SO₂ — so such grapes need a higher dose.
Potassium metabisulfite ($K_2S_2O_5$) is a white powder that releases SO₂ when dissolved in water or must. From the molar masses, 1 g of $K_2S_2O_5$ gives ≈ 0.576 g of SO₂. The powder is first dissolved in a small amount of must or water and then mixed evenly through the whole volume.
Sulfurous acid solution (usually 5–6% SO₂) is a ready-made solution; 1 mL of a 6% solution contains ≈ 60 mg of SO₂. It is convenient, but over time SO₂ evaporates and the strength falls — the solution must be kept tightly closed.
A sulfur wick is not used for dosing must — it is for sulfuring empty vessels (qvevri, barrels). On burning, $S + O_2 \rightarrow SO_2$: 1 g of sulfur gives 2 g of SO₂, which destroys microflora on the vessel walls. The practical rate is about 2–3 g of sulfur per 100 L of vessel volume.
A small spherical particle in a viscous liquid settles at constant speed once gravity (minus buoyancy) balances drag:
$$v = \frac{2\,r^2\,(\rho_p - \rho)\,g}{9\,\mu}$$
where $r$ is the particle radius, $\rho_p$ and $\rho$ the densities of particle and must, and $\mu$ the must viscosity. Velocity is proportional to the square of the radius — a particle twice as large settles four times as fast. The density difference in must is very small (0.02–0.05 g/cm³), so individual cell fragments hardly settle at all; settling is only possible when particles join into aggregates of 30–100 µm.
Must viscosity rises with sugar and with falling temperature. The calculator uses a sucrose-solution approximation: $\mu_{20} = \exp(0.0229B + 0.000527B^2)$ mPa·s, with temperature entering in Arrhenius form: $\mu(T) = \mu_{20}\exp\left[2320\left(\frac{1}{T} - \frac{1}{293.15}\right)\right]$. The viscosity of a 22 Brix must is ≈ 2.1 mPa·s at 20 °C and ≈ 2.8 at 10 °C.
Grape pectin raises viscosity and keeps particles in a colloidal state. A pectolytic enzyme breaks the pectin chains, lowers viscosity and lets the particles join into aggregates. This speeds settling several-fold.
The paradox of cold settling: in the cold viscosity is higher and particles fall more slowly, yet clarification is still better. The reason is that cold delays wild-yeast fermentation. At 20 °C fermentation can start within a day, and CO₂ bubbles lift the sediment; at 8–10 °C fermentation is delayed by 2–3 days, leaving the sediment time to settle. The calculator compares both times: the clearing time and the onset of wild fermentation ($\approx 24$ h at 20 °C, $Q_{10} = 2$).
Active dry yeast contains about $(1-2)\cdot 10^{10}$ viable cells per gram. The usual inoculation rate is 20–30 g/hL, which gives 2–6 million cells per mL of must:
$$N = \frac{q}{100}\cdot c\cdot 10^{-3}\quad (\text{cells/mL})$$
where $q$ is the rate (g/hL) and $c$ the cells per gram. Dry yeast is first rehydrated: in 10 times its mass of warm water (35–40 °C) for 15–20 minutes. Adding it straight into cold must gives the cells a thermal shock, so the temperature is lowered in steps — adding must each time, so that the difference never exceeds 10 °C.
Yeast needs nitrogen to make proteins and multiply. It can only use assimilable nitrogen — ammonium and free amino acids (except proline). A nitrogen shortage slows or stops fermentation and causes hydrogen sulfide (H₂S, the "rotten egg" smell). The need rises with sugar; a common guide is:
$$YAN_{\text{need}} \approx 200 + 25\,(B - 21)\ \ \text{mg N/L}$$
A shortfall is made up with diammonium phosphate (DAP, $(NH_4)_2HPO_4$): 1 g/L of DAP gives 212 mg N/L. The EU allows ammonium salts up to 1 g/L. DAP is best added in two parts — at the start of fermentation and after a third of the sugar has fermented.
| Quantity | Typical value | Note |
|---|---|---|
| SO₂ in healthy must | 30 – 50 mg/L | a little more for white, less for red |
| SO₂ in damaged grapes | 50 – 70 mg/L | hail, insects |
| SO₂ in rotten grapes | 70 – 100 mg/L | laccase is resistant |
| pKa of molecular SO₂ | 1.81 | 20 °C |
| SO₂ in potassium metabisulfite | 57.6% | 1 g → 0.576 g SO₂ |
| Sulfur wick | 2 – 3 g / 100 L | for empty vessels |
| EU limit (dry) | red 150, white 200 mg/L | total SO₂ |
| Viscosity at 22 Brix, 20 °C | ≈ 2.1 mPa·s | water ≈ 1.0 |
| Yeast rate | 20 – 30 g/hL | 2–6 million cells/mL |
| Nitrogen need | 140 – 350 mg N/L | by sugar level |
| DAP | 212 mg N per 1 g | limit 1 g/L |
The previous pages are the foundation of this one: "Acids, pH and the buffer system" computes the must pH that sets SO₂ effectiveness; "Sugars and must density" gives the sugar and density that determine viscosity, settling and nitrogen need; "Harvest and must yield" gives the volumes for which the doses are calculated.
Chapter III (Alcoholic fermentation) starts where this page ends: with yeast growth kinetics, nitrogen uptake and fermentation temperature. Chapter IV (Wine chemistry and stabilisation) deals with SO₂ in finished wine — how it is consumed through oxidation and ageing.
Preparing the qvevri is an inseparable part of Georgian winemaking. The qvevri is washed with water and lime, scrubbed with a brush made from cherry bark, and finally a sulfur wick is burned inside — this cleans the walls of the qvevri and suppresses the microflora left in its pores.
Traditional qvevri wine is often made without sulfiting, by wild (spontaneous) fermentation: the must is fermented by the microflora living on the grape skins and in the marani. During fermentation on the pomace, phenolics, CO₂ and a quick start to fermentation give the must some protection. This approach gives the wine an individual character, but it demands healthy grapes, cleanliness and close control of fermentation — with rotten grapes, or high pH in a hot year, the risk rises sharply.
Sulfiting and a pure yeast strain give a predictable, "clean" fermentation, which is especially valuable for large volumes and damaged grapes. Both approaches are legitimate — the choice depends on the winemaker's aim, the condition of the grapes and the willingness to accept risk.
| Term | Definition |
|---|---|
| Sulfiting | Adding SO₂ to must or wine, or sulfuring a vessel |
| Free SO₂ | SO₂ not bound to carbonyls (molecular + bisulfite) |
| Bound SO₂ | SO₂ bound to carbonyl compounds (acetaldehyde etc.) |
| Molecular SO₂ | The antimicrobially active form of SO₂ |
| Total SO₂ | The sum of free and bound SO₂ |
| Potassium metabisulfite | K₂S₂O₅ — a solid source of SO₂ |
| Must settling | Letting particles settle out before fermentation |
| Pectolytic enzyme | An enzyme that breaks down pectin and speeds settling |
| Pure yeast strain | Yeast of a selected strain, usually Saccharomyces cerevisiae |
| Yeast rehydration | Reviving dry yeast in warm water before inoculation |
| Assimilable nitrogen | Nitrogen available to yeast (ammonium, amino acids) |
| Diammonium phosphate (DAP) | A nitrogen nutrient for yeast |
| Spontaneous fermentation | Fermentation by wild microflora, without inoculation |
1000 litres of healthy Rkatsiteli must, pH 3.40. The winemaker adds 40 mg/L of SO₂; the bound fraction is 35%, so free SO₂ is 26 mg/L and molecular SO₂ 0.65 mg/L — only 2.5% of the free. That is enough to suppress the wild microflora. For a target of 0.5 mg/L molecular SO₂, 31 mg/L would have sufficed. The same must at pH 3.80 would need about 76 mg/L for the same protection. A 40 mg/L dose takes 69 g of potassium metabisulfite, or 667 mL of 6% solution. To sulfur a 2000-litre qvevri before filling, about 50 g of sulfur wick is burned.
The must (22 Brix) settles in a tank 1.5 m high, with a pectolytic enzyme. At 10 °C the viscosity is 2.82 mPa·s, 40 µm aggregates settle at 3.3 cm/h, and the must clears in about 45 hours. At this temperature wild fermentation would start in about 48 hours, so clarification finishes first. At 20 °C clarification is faster (≈ 34 h), but fermentation starts in ≈ 24 hours — risky. Without enzyme, at 10 °C, clearing would take about 175 hours.
The settled must is inoculated with a pure yeast strain at 25 g/hL: 250 g of yeast, ≈ 5 million cells per mL of must. Rehydration needs 2.5 L of water at 35–40 °C; the must is at 16 °C, so the temperature must be lowered in two steps. The must is at 24 Brix and its measured assimilable nitrogen is 150 mg N/L. The need is ≈ 275 mg N/L, so 0.59 g/L of DAP — 590 g in all — should be added, which is within the EU limit.
1. In the "Sulfiting" mode choose the grape condition — the dose and bound fraction are set automatically. Enter the must pH (if you don't know it, compute it on the "Acids, pH and the buffer system" page) and the volume. The "Amount required" card tells you how many grams of metabisulfite or millilitres of solution you need.
2. Look at the molecular SO₂ card. If it is red, protection at this pH is weak — either raise the dose or lower the pH first. The "Dose for the target" card shows exactly how much SO₂ is needed for the target molecular level.
3. To sulfur a qvevri or barrel, enter the vessel volume — you get the amount of sulfur wick. The wick is burned only in an empty, clean vessel, before filling.
4. For white wine use the "Must settling" mode. Enter the temperature and the tank height. If the card warns you in red, chill the must or add enzyme — settling must finish before wild fermentation starts.
5. In the "Yeast and nitrogen" mode enter the volume, yeast rate, sugar and measured nitrogen. You get the amount of yeast, the rehydration water and how much diammonium phosphate to add.
1. The bound fraction is the most uncertain parameter. If you can, measure free SO₂ 1–2 hours after the addition by the aeration-oxidation method (or iodometry) and adjust the parameter so the model matches your result.
2. pKa = 1.81 applies at 20 °C and low ionic strength. As temperature and alcohol rise, pKa falls and the molecular share rises, so SO₂ is more effective (and more noticeable on the nose) in warm wine. At the must stage this effect is small.
3. Stokes' law applies to a single spherical particle. In real must particle sizes are distributed, and at high concentration settling slows through mutual hindrance. Use the "effective diameter" as a calibration parameter: measure how long your tank takes to clear and adjust the diameter.
4. The onset time of wild fermentation depends strongly on the size of the microbial population, on SO₂ and on oxygen. In sulfited must it is longer, so the calculator's curve is a conservative estimate for unsulfited must.
5. The nitrogen-need formula is a guide. The real need depends on the strain, temperature and oxygen; some strains need much less. DAP supplies only inorganic nitrogen — with high sugar or difficult conditions it is better to make up part with organic nutrients (yeast extract, amino acids) and add the dose in two parts.
1. 40 mg/L of SO₂ was added to the must. Does this kill all wild yeast?
Sulfiting is selection, not sterilisation. Non-Saccharomyces yeasts and acetic bacteria are sensitive to SO₂; Saccharomyces cerevisiae is tolerant, and a pure strain inoculated in sufficient numbers soon dominates.
2. One must has pH 3.1, another 3.7. Roughly how many times more free SO₂ does the second need for the same molecular SO₂?
Molecular share = 1/(1 + 10^(pH − 1.81)). At pH 3.1 ≈ 4.9%, at pH 3.7 ≈ 1.3% — almost a fourfold difference. That is why controlling pH is a precondition for sulfiting.
3. Why does must settle better when cold, even though cold liquid is more viscous and particles fall more slowly?
At 20 °C wild fermentation can start within a day, and CO₂ bubbles lift the sediment. At 8–10 °C fermentation is delayed by 2–3 days, so the sediment settles, if more slowly.
4. A must has 24 Brix and measured assimilable nitrogen of 150 mg N/L. Roughly how much DAP does it need?
Need ≈ 200 + 25·(24 − 21) = 275 mg N/L; shortfall 125 mg N/L. DAP gives 212 mg N/L per 1 g/L, so 125/212 ≈ 0.59 g/L.