Winemaking Laboratory
Grape chemistry and harvest / Acids, pH and the buffer system

Chapter II · Grape chemistry and harvest

Acids, pH and the buffer system

Why pH and total acidity do not move together, in what forms tartaric and malic acid exist in must, how potassium acts, and how much acid or carbonate is needed to raise or lower acidity.

Must composition

The model solves the charge balance: all acid anions = H⁺ + K⁺ + other cations. Activity coefficients from the Davies equation; phosphate is fixed (2 mmol/L). Amino acids and phenolics are not included, so the real buffer capacity is 10–20% higher.

pH—20 °C
Total acidity—g/L as tartaric · pH 7.0
Same as sulfuric—g/L H₂SO₄
Buffer capacity—meq/L · pH⁻¹
Bitartrate HT⁻—% of tartaric acid
Free H₂T—% of tartaric acid
Malic share of acidity—%
KHT product [K⁺]·[HT⁻]—· 10⁻⁴ (mol/L)²

Adjustment

The calculator does not include later precipitation of cream of tartar (KHT) or calcium tartrate: after a tartaric addition part of the acidity is lost during cold stabilisation, and after CaCO₃ calcium keeps precipitating for months. Trial the dose on 1 litre first.

New pH—
New acidity—g/L
Total required—
Dose for the target pH—

Titration result

CO₂ must first be removed from the sample (warming or vacuum), otherwise carbonic acid is titrated too and acidity comes out too high. The curve is built for the composition entered in the first mode.

Acidity as tartaric—g/L
Acidity as sulfuric—g/L
Equivalents—meq/L
Error of one drop (0.05 mL)—g/L

Intuition

Acidity is the backbone of wine: it gives freshness, balances sugar and alcohol, protects the wine from microbes and determines how long it will age. But "acidity" means two different things, and telling them apart is one of a winemaker's key skills.

Total acidity (titratable acidity) tells us how much acid is in the must — the number of protons that can be neutralised with a base. pH tells us how strongly that acid acts — how many free hydrogen ions are in solution at this moment. Perceived sourness on the palate is driven more by total acidity, while chemical and microbial stability, colour and the effectiveness of sulfur dioxide are governed by pH.

Two musts can have the same total acidity and completely different pH. The reason is the buffer system: grape acids are weak, part of them exists as potassium salts, and this mixture resists changes in pH. On this page we build that system quantitatively.

History

Cream of tartar — the crystals deposited on the walls of barrels and qvevri — has been known since antiquity. In 1769 the Swedish chemist Carl Wilhelm Scheele first isolated tartaric acid from it, and in 1785 malic acid from apple juice. In 1848 the young Louis Pasteur, studying crystals of tartaric acid salts, discovered molecular asymmetry (chirality) — one of the fundamental discoveries of chemistry began with cream of tartar.

The concept of pH was introduced in 1909 by the Danish chemist Søren Sørensen at the Carlsberg Laboratory in Copenhagen, during his research on enzymes. Titration analysis of wine was already routine in the 19th century, while systematic pH measurement became established in winemaking from the mid-20th century, with the spread of electronic pH meters.

At the end of the 19th century the Swiss researcher Hermann Müller-Thurgau linked the spontaneous decrease in acidity of young wine to bacterial activity. This began the study of malolactic fermentation, which today is part of almost every red winemaking technology.

Must acids

More than 90% of grape acidity comes from two diprotic acids. Tartaric acid is almost unique to the vine — it is rare in other fruit. Its amount in the berry is formed before véraison and then hardly changes: during ripening its concentration falls only through berry growth (dilution). Must usually contains 4–8 g/L. Tartaric is the strongest of the grape acids (pKa₁ = 3.04) and is microbially stable.

Malic acid, by contrast, is a respiratory substrate: during ripening the vine "burns" it, and the warmer the nights, the faster. Grapes from a cool climate may therefore contain 4–7 g/L, while in a hot year it falls to 1–2 g/L — the process described with Q₁₀ = 2 on the "Ripening dynamics" page of Chapter I. Malic acid is available to bacteria and is converted to lactic acid during malolactic fermentation.

Citric acid (0.1–0.5 g/L), phosphoric acid and amino acids are present in small amounts; after fermentation succinic, lactic and acetic acids are added. The main cation is potassium (1000–2500 mg/L): the roots take it up from the soil, it is concentrated mostly in the skins, and it passes into the must on crushing, and especially during maceration (time on the pomace).

pH, dissociation and acid forms

A diprotic acid $H_2A$ dissociates in two steps: $H_2A \rightleftharpoons H^+ + HA^-$ (constant $K_1$) and $HA^- \rightleftharpoons H^+ + A^{2-}$ ($K_2$). At any pH the acid is distributed among three forms, and their fractions depend only on pH and the pKa values:

$$\alpha_0 = \frac{1}{D},\quad \alpha_1 = \frac{K_1/[H^+]}{D},\quad \alpha_2 = \frac{K_1K_2/[H^+]^2}{D},\qquad D = 1 + \frac{K_1}{[H^+]} + \frac{K_1K_2}{[H^+]^2}$$

When $pH = pK_a$, the two neighbouring forms are equal (Henderson–Hasselbalch equation: $pH = pK_a + \log([A^-]/[HA])$). At the usual must pH (3.0–3.8) tartaric acid exists mainly as bitartrate ($HT^-$), while malic acid is a mixture of the free acid and $HM^-$. The bitartrate fraction peaks around pH 3.6–3.7 — which is why cream of tartar ($KHT$) precipitation is most intense at this pH.

Mathematics: the charge balance

The solution is electrically neutral, so the sum of positive charges equals the sum of negative ones:

$$[H^+] + [K^+] + \sum[M^{n+}]\,n = \sum_{\text{acids}} C_i\,(\alpha_{1,i} + 2\alpha_{2,i} + 3\alpha_{3,i}) + [OH^-]$$

On the left are the hydrogen ion, potassium and other cations (Ca, Mg, Na, in equivalents); on the right, all the acid anions. The only unknown in this equation is $[H^+]$; the right-hand side increases monotonically with pH, so the solution is unique and the calculator finds it by bisection.

Activity. A pH meter measures activity $a_{H^+} = \gamma\,[H^+]$, not concentration. The ionic strength of must is $I \approx 0.04–0.07$ mol/L, which lowers activity coefficients noticeably. The calculator computes them with the Davies equation:

$$\log_{10}\gamma_z = -0.51\,z^2\left(\frac{\sqrt I}{1+\sqrt I} - 0.3\,I\right)$$

and refines ionic strength and pH iteratively. For a doubly charged ion ($z = 2$) $\gamma \approx 0.4$, so the thermodynamic pKa₂ is "effectively" 0.4–0.5 units lower in must.

Total acidity is the amount of base that takes the must from its initial pH to the titration end point (pH 7.0 or 8.2). Converting equivalents to grams: $TA_{\text{tartaric}} = 0.07504 \cdot TA_{\text{meq/L}}$, $TA_{\text{sulfuric}} = 0.04904 \cdot TA_{\text{meq/L}}$, i.e. $TA_{\text{sulfuric}} = 0.654\,TA_{\text{tartaric}}$. In a laboratory titration:

$$TA\ (\text{g/L as tartaric}) = \frac{V_{NaOH}\cdot N \cdot 75.04}{V_{\text{sample}}}$$

Buffer capacity $\beta = dC_B/d\,pH$ — how many milliequivalents of base are needed to raise pH by one unit. For must it is usually 30–60 meq/L·pH. The larger β, the less pH changes when acid or carbonate is added.

Key quantities

QuantitySymbolUnitTypical value in must
pHpH—3.0 – 3.8
Total (titratable) acidityTAg/L as tartaric4 – 10
Tartaric acid—g/L4 – 8
Malic acid—g/L1 – 7
Citric acid—g/L0.1 – 0.5
PotassiumK⁺mg/L1000 – 2500
Ionic strengthImol/L0.04 – 0.07
Buffer capacityβmeq/L·pH⁻¹30 – 60

Acid dissociation constants (25 °C, thermodynamic)

AcidProtonspKa₁pKa₂pKa₃Molar mass (g/mol)
Tartaric23.044.37—150.09
Malic23.405.20—134.09
Citric33.134.766.40192.12
Lactic13.86——90.08
Phosphoric32.157.2012.3597.99

Potassium and cream of tartar

Potassium is the main "enemy" of pH. Each potassium ion replaces one proton of an acid, so more potassium raises pH and lowers total acidity — as if base had been added to the must. Must contains more potassium the more fertile the soil, the greater the vine's vigour, the more shaded the clusters and the longer the must stays on the pomace.

Potassium bitartrate ($KHT$, cream of tartar) is poorly soluble, and its solubility drops sharply as alcohol rises and temperature falls. After fermentation wine is almost always supersaturated, and crystals precipitate in winter or during cold stabilisation. This shifts pH: below pH 3.65, cream of tartar precipitation lowers pH (more base, $K^+$, leaves the solution than acid), above 3.65 it raises it. The calculator shows the concentration product $[K^+][HT^-]$ as a guide; the real saturation limit in wine is higher because of complexation and colloids and is determined with a conductivity test.

Raising / lowering acidity

Acidification is carried out in hot years or when pH is too high. The main agent is L-tartaric acid: 1 g/L initially raises total acidity by exactly 1 g/L and lowers pH by 0.1–0.2, depending on buffer capacity. Part of it later precipitates as cream of tartar, so the final effect is smaller. Lactic acid is monoprotic and weak: more is needed for the same pH effect, but it does not precipitate. Under EU rules, acidification of must is allowed up to 1.5 g/L (as tartaric acid, i.e. 20 meq/L) and of wine up to 2.5 g/L.

Deacidification is carried out in cool, acidic years. The stoichiometry is based on equivalents: $CaCO_3$ (100 g/mol) neutralises two equivalents, so lowering acidity by 1 g/L takes 0.67 g/L; $KHCO_3$ is monovalent and takes 1.33 g/L. In both cases CO₂ is released. Calcium precipitates as calcium tartrate over months, while potassium raises pH further. In the EU, deacidification of wine is allowed up to 1 g/L (as tartaric acid).

Malolactic fermentation is biological deacidification: $C_4H_6O_5 \rightarrow C_3H_6O_3 + CO_2$. Diprotic malic acid becomes monoprotic lactic acid, so converting 1 g/L of malic acid lowers acidity by about 0.56 g/L (as tartaric) and raises pH by 0.1–0.3. The taste softens; the acidity no longer "bites".

Connections

The previous page, "Sugars and must density", is the other side of the balance: sugar and total acidity together define the style of a wine, and on the "Ripening dynamics" page of Chapter I both change over time, with pH computed by a simplified formula. Here the same pH is solved with an exact charge balance. The next page, "Phenolic compounds", shows why pH is critical for colour: the share of the red form of anthocyanins falls rapidly as pH rises.

In Chapter III, pH determines what share of sulfur dioxide is in the active molecular form (the difference between pH 3.2 and 3.8 is almost fourfold) and how safe the wine is from bacteria. Malolactic fermentation is treated there in detail.

Application — Georgian practice

Georgian varieties differ sharply in acidity. The white varieties of western Georgia (Tsitska, Tsolikouri) keep high acidity — which is exactly why Tsitska suits sparkling wine. Rkatsiteli holds its acidity well even in hot Kakheti, and Saperavi is a high-acid red variety, which explains its long ageing potential.

In the Kakhetian tradition the must stays on the pomace for months. Potassium passes from the skins, so the pH of qvevri wine is usually higher, and total acidity lower, than in a wine of the same variety made the European way. Add to this the malolactic fermentation that runs spontaneously in the qvevri over winter and the deposition of cream of tartar on the qvevri walls. For the winemaker this means that in a hot year pH must be controlled before the wine is set: after maceration, correction is harder.

Glossary

TermDefinition
pHNegative decimal logarithm of hydrogen ion activity
Total acidity (titratable acidity)Sum of all titratable acids, expressed as tartaric or sulfuric acid
Dissociation constant (pKa)The pH at which two neighbouring forms of an acid are equal
Buffer capacityThe amount of base that raises pH by one unit
BitartrateThe singly dissociated form of tartaric acid, HT⁻
Potassium bitartrate (cream of tartar)KHT — the poorly soluble potassium salt of tartaric acid
Milliequivalent (meq)The number of neutralisable protons, in millimoles
Ionic strengthA combined measure of ion concentration and charge in a solution
Activity coefficientRatio of an ion's "effective" concentration to its actual concentration
Raising / lowering acidityCorrecting acidity by adding acid or carbonate
Malolactic fermentationConversion of malic acid to lactic acid by bacteria
Cold stabilisationChilling wine to precipitate cream of tartar in advance

Practical use in calculations

It is a hot year in Kakheti. Analysis of a Rkatsiteli must shows tartaric acid 5.0 g/L, malic acid 2.0 g/L, citric acid 0.3 g/L and potassium 1800 mg/L. From the charge balance the calculator computes pH 3.68 and total acidity 3.9 g/L. For a qvevri wine that will take up even more potassium on the pomace, that pH is dangerously high and the acidity low. Buffer capacity is about 40 meq/L·pH. The winemaker chooses pH 3.50 as the target. The acidity adjustment mode shows that this needs about 1.07 g/L of tartaric acid. That is within the EU limit (1.5 g/L) and raises total acidity to 4.9 g/L; for 5000 litres it is 5.35 kg of tartaric acid. In the laboratory, titrating a 10 mL sample with 0.1 N NaOH uses about 5.2 mL before the addition. For comparison, a normal-year must (tartaric 6, malic 3 g/L, potassium 1600 mg/L) has pH 3.31 and acidity 6.4 g/L. Full malolactic fermentation would raise its pH to 3.44 and lower its acidity to 4.7 g/L.

For beginners — from analysis to decision

1. Open the "Calculator", mode "Acids and pH". Enter the laboratory results: tartaric acid, malic acid and potassium. If you do not know citric acid and the other cations, leave the defaults.

2. The cards show the computed pH, total acidity as tartaric and as sulfuric acid (older Georgian and Russian literature often used sulfuric), and the buffer capacity. If you have also measured the must's pH, compare: a difference greater than 0.1 means something is missing from the analysis, usually potassium.

3. The first chart shows in what form tartaric and malic acid exist at your pH. The second shows how potassium changes pH and acidity. Try moving the potassium slider and see how sensitive pH is.

4. If pH is too high (above 3.5 for white, above 3.7 for red), switch to the "Acidity adjustment" mode. Choose the agent and target pH and enter the volume. The calculator tells you how many kilograms are needed and whether the dose is within the EU limit.

5. The "Titration" mode converts a laboratory result: enter the sample volume, NaOH normality and millilitres used. You get acidity in all standard units.

For professionals — model limits and accuracy

1. The model solves the exact charge balance with thermodynamic pKa values and Davies activity coefficients. It does not include amino acids, phenolics, proteins or succinic acid, so the real buffer capacity is 10–20% higher and the pH change on addition slightly smaller. In wine, ethanol raises the pKa values by 0.1–0.2; this effect is not included either.

2. The "other cations" parameter is the model's calibration knob. If you have measured pH, total acidity, tartaric acid, malic acid and potassium on one sample, adjust this parameter until the computed pH matches the measured one. The adjustment forecast will then be much more accurate.

3. The real effect of a tartaric acid addition is smaller than computed, because part of it precipitates as cream of tartar. In a wine below pH 3.65 this precipitation lowers pH further; above it, it partly reverses the change. The final effect only shows after cold stabilisation, so always trial the dose on a 1-litre sample first.

4. The two titration end points give different results: titrating to pH 8.2 (phenolphthalein) reads 0.05–0.3 g/L higher than titrating to pH 7.0 (the OIV method). The difference comes from phosphate, phenolics and amino acids. When comparing results from different laboratories, always check the method.

5. In red must the colour masks the indicator change, so titration with a pH meter is better. CO₂ must be removed from the sample (brief boiling or vacuum), because carbonic acid raises acidity by 0.2–0.5 g/L. NaOH solution absorbs CO₂ from the air, so its normality should be checked regularly against potassium hydrogen phthalate.

1. Two musts have the same total acidity (6 g/L), but one has pH 3.2 and the other 3.6. What is the main reason for the difference?

Total acidity counts every titratable proton, pH only the free H⁺. Potassium turns part of the acids into salts — pH rises while titratable acidity changes much less.

2. How much CaCO₃ lowers acidity by 1 g/L (as tartaric)?

One mole of CaCO₃ (100 g) neutralises two equivalents, i.e. 50 g/eq; tartaric acid is 75 g/eq, so 50/75 ≈ 0.67. KHCO₃ is monovalent (100 g/eq) and needs 1.33 g/L.

3. Why does malolactic fermentation lower acidity?

L-malic acid → L-lactic acid + CO₂. Converting 1 g/L of malic acid lowers acidity by about 0.56 g/L (as tartaric) and raises pH by 0.1–0.3.

4. A must has pH 3.3. Which form of tartaric acid dominates?

Between pH 3.04 (pKa₁) and 4.37 (pKa₂) the main form is HT⁻. It is the one that forms cream of tartar (KHT) with potassium, and precipitation is strongest around pH 3.6–3.7.