Winemaking Laboratory
Grape chemistry and harvest / Sugars and must density

Chapter II · Grape chemistry and harvest

Sugars and must density

What the refractometer and hydrometer measure, how to convert between Brix, Oechsle, Baumé and specific gravity, how much alcohol to expect, and how many berries to sample before trusting the measurement.

Measurement

Both instruments are calibrated at 20 °C. Refractometer correction: 0.075 Brix/°C; hydrometer correction: from the temperature dependence of water density. For a digital refractometer with automatic correction (ATC), set the temperature to 20 °C.

Sugar content—°Brix
Specific gravity—20/20 °C
Oechsle—°Oe
Baumé—°Bé
Sugar—g/L
Potential alcohol (legal)—% vol · 16.83 g/L
Expected alcohol—
Theoretical maximum—% vol · 15.43 g/L

Two musts

Must A (main)

Must B (to add)

Calculated by mass balance (Brix is a mass quantity): m = V · SG. The last card shows how far a simple volume-based "Pearson square" would be off.

Must B needed—L
Blend volume—L
If all of B is added—°Brix
Volume-based error—Lcompared with the mass balance

Sample size and precision

n = (z·σ/E)². The formula assumes berries are taken at random from different parts of the plot — berries from one cluster resemble each other and reduce the real precision.

Required sample—berries
Precision of your sample—°Brix
Error in alcohol—% vol
For comparison: ±0.5 Brix—berries

Intuition

Grape must is, first of all, a solution of sugar in water: a litre of ripe must contains 180–260 grams of sugar, mostly glucose and fructose. This sugar is the only source of the future wine's alcohol, so measuring it accurately is the winemaker's first and most important number — it determines the wine's body, its balance, the length of fermentation and sometimes even its legal category.

Sugar cannot be weighed directly in the vineyard, so we measure it indirectly — through a physical property of the solution. Sugar increases the density of the solution (must is heavier than water) and its refractive index (light is bent more strongly in must). The first is measured by the hydrometer, the second by the refractometer. Both instruments actually measure the total of soluble solids, of which more than 90% is sugar and the rest acids, minerals, phenolics and nitrogenous compounds.

On this page we will learn what each scale measures (Brix, Oechsle, Baumé, specific gravity), how to convert between them, why the measurement needs temperature correction, how to predict alcohol and how many berries we must sample before we can trust the result.

History

The measurement of must density began in the 18th century. In 1768 the French chemist Antoine Baumé created a hydrometer scale that is still used in France — its convenience is that a degree Baumé roughly corresponds to the percentage of alcohol in the future wine. In the 1830s the German Ferdinand Oechsle introduced a scale that simply shows how many grams heavier a litre of must is than a litre of water; Oechsle degrees remain the basis of German and Swiss wine law.

In 1843 the Bohemian Karl Balling compiled tables relating the density of sucrose solutions to their mass percentage; in 1854 Adolf Brix refined them. Thus the Brix scale was born — grams of sucrose per 100 grams of solution — which is today the most widely used in the world. In the 20th century the hand refractometer, which needs only a few drops of juice, almost completely drove the hydrometer out of the vineyard; in Georgian practice its reading is referred to as "sugar percent".

The quantitative law of the conversion of sugar into alcohol was formulated by Joseph Gay-Lussac in 1815, and in the 1850s–60s Louis Pasteur showed that real fermentation does not fully follow this theoretical equation — part of the sugar is spent on yeast biomass, glycerol and other by-products.

The sugars of must

Glucose and fructose make up 98–99% of the sugars in ripe grapes, in roughly equal amounts. Sucrose formed by photosynthesis in the leaf is transported through the phloem into the berry, where the enzyme invertase splits it into glucose and fructose; sucrose is therefore present only in traces in ripe Vitis vinifera grapes (hybrid varieties contain more). Around véraison glucose slightly predominates; at full ripeness the glucose/fructose ratio approaches 0.9–1.0, and in over-ripe grapes fructose predominates.

This matters for two reasons. First, fructose is almost twice as sweet as glucose, so the perceived sweetness of a semi-sweet wine depends on the composition of its residual sugar. Second, yeast preferentially consumes glucose, so the sugar left at the end of fermentation is mainly fructose; this explains why a wine that has not fermented completely can taste noticeably sweet even with a small amount of sugar.

The physics of measurement: hydrometer and refractometer

The hydrometer relies on Archimedes' principle: the float sinks into the solution until the weight of the liquid it displaces equals its own weight. The denser the must, the less the instrument sinks, and the scale on its stem shows the specific gravity (relative density) or the Oechsle, Baumé or Brix degree directly. A hydrometer needs 100–250 mL of must, the sample must be clean (particles and gas bubbles distort the result), and the instrument is calibrated at a particular temperature — usually 20 °C.

The refractometer measures the refractive index $n$ — how much light is slowed down in the solution. For pure water $n_D = 1.3330$; for a 20% sucrose solution it is about 1.3639. The instrument's scale is calibrated on sucrose solutions and reads Brix directly. Its advantage is that a few drops are enough; its drawbacks are sensitivity to temperature and the fact that once fermentation has started, alcohol distorts the reading — so the refractometer is no longer usable in fermenting must, where only the hydrometer works.

Temperature correction. In warm must the hydrometer sinks too far (the liquid expands and its density falls) and underestimates sugar; the refractometer also reads low on a warm sample — by about 0.07–0.08 Brix for every degree above 20 °C. On a hot harvest day, with a sample at 30 °C, an uncorrected result is wrong by almost one Brix. Modern digital refractometers make this correction automatically (ATC).

Mathematics

Definition of Brix — the mass fraction of sucrose in percent: $B = 100 \cdot m_{\text{sugar}} / m_{\text{solution}}$.

The relationship between Brix and specific gravity (20 °C, an approximation fitted to sucrose solution tables):

$$SG = 1 + \frac{B}{258.6 - 0.8796\,B}, \qquad B = \frac{258.6\,(SG-1)}{1 + 0.8796\,(SG-1)}$$

Oechsle and Baumé follow directly from specific gravity: $°Oe = 1000\,(SG - 1)$, $°Bé = 145\,(1 - 1/SG)$.

Sugar concentration per litre: the mass of soluble solids in one litre is $10 \cdot B \cdot SG$ grams, of which about 90% is sugar: $S \approx 10\,B\,SG\,(1 - f_{ns})$, where $f_{ns} \approx 0.10$ is the fraction of non-sugars (acids, minerals, phenolics). For a 22 Brix must, $SG \approx 1.092$ and $S \approx 216$ g/L.

Stoichiometry of alcoholic fermentation (Gay-Lussac): $C_6H_{12}O_6 \rightarrow 2\,C_2H_5OH + 2\,CO_2$. 180 g of hexose theoretically yields 92 g of ethanol (51.1% by mass). One percent by volume of alcohol is 7.89 g of ethanol per litre, so theoretically it requires 15.4 g/L of sugar. In reality 5–10% of the sugar is spent on yeast biomass, glycerol (4–5 g/L), succinic acid and other by-products, and some ethanol evaporates with the CO₂. The practical coefficient is therefore 16.5–18 g/L; under EU law potential alcohol is calculated at 16.83 g/L:

$$PA\,(\%\,vol) = \frac{S\,(g/L)}{16.83}$$

Blending two musts. Brix is a mass quantity, so the exact balance is written by mass: $m_A B_A + m_B B_B = (m_A + m_B)\,B_T$, where $m = V \cdot SG$. From this the required mass of the second must is $m_B = m_A\,(B_T - B_A)/(B_B - B_T)$. This is the well-known "Pearson square", written in a physically correct form.

Sample size. Sugar content varies considerably between individual berries (standard deviation $\sigma \approx 1.5–2.5$ Brix). The standard error of the mean of $n$ berries is $\sigma/\sqrt{n}$, so to reach a desired precision $E$ with 95% confidence one needs

$$n = \left(\frac{1.96\,\sigma}{E}\right)^2$$

berries. For $\sigma = 2$ Brix and $E = \pm 0.3$ Brix this is about 170 berries — which is where the practical rule of "200 berries" comes from.

Key quantities

QuantitySymbolUnitTypical value at harvest
Sugar content (Brix)B°Brix (g/100 g)18 – 26
Specific gravity (relative density)SG—1.074 – 1.110
Oechsle degree°Oe—74 – 110
Baumé degree°Bé—10 – 14.4
Sugar concentrationSg/L175 – 260
Potential alcoholPA% vol10.3 – 15.4
Refractive index$n_D$—1.360 – 1.375
Glucose / fructoseG/F—0.9 – 1.0
Sugar per 1% alcohol—g/L16.5 – 18 (16.83 by law)
Berry-to-berry spreadσ°Brix1.5 – 2.5

Comparison of scales

BrixSpecific gravityOechsleBauméSugar (g/L)PA (% vol)
181.0747410.017410.3
201.0838311.119511.6
211.0888811.720612.2
221.0929212.221612.8
241.10110113.323814.1
261.11011014.426015.4

Connections

Sugar content is the main variable of the "Ripening dynamics" page in Chapter I — there it changes over time, here we measure and analyse it at a single moment. The next page, "Acids, pH and the buffer system", is the other half of the picture: the balance between sugar and total acidity (titratable acidity) determines the wine's style. "Harvest and must yield" shows how sugar content differs between free-run and press juice.

In Chapter III (alcoholic fermentation) the sugar concentration calculated here is the initial condition for fermentation kinetics, and the hydrometer is the main tool for monitoring fermentation. There we will also discuss sugar correction (chaptalisation) and why very high sugar can stop fermentation.

Application — Georgian practice

In the Georgian marani, sugar is traditionally measured with a refractometer in the vineyard and, once the must is in the qvevri, with a hydrometer, because the refractometer is unreliable in must fermenting together with the pomace. The technical specifications of wines with a protected designation of origin (PDO) also set a minimum sugar content for the grapes, so an accurate, temperature-corrected measurement is also a legal requirement.

For naturally semi-sweet wines (Khvanchkara, Kindzmarauli, Tvishi) sugar content is especially critical: the grapes are picked at high sugar and fermentation is stopped while residual sugar is retained. Here the winemaker needs to calculate in advance what alcohol and what residual sugar he will obtain — exactly with the calculator on this page.

Glossary

TermDefinition
MustThe juice of crushed grapes before fermentation
Brix (°Brix)Grams of sucrose per 100 grams of solution; in Georgian practice "sugar percent"
Specific gravity (relative density)The ratio of the density of must to that of water at the same temperature
Oechsle degreeHow many grams heavier a litre of must is than a litre of water
Baumé degreeA hydrometer scale that roughly corresponds to potential alcohol
HydrometerA floating instrument that measures density by Archimedes' principle
RefractometerAn instrument that measures soluble solids through the refractive index
Soluble solidsAll solid compounds dissolved in the must (sugars, acids, minerals, phenolics)
InvertaseThe enzyme that splits sucrose into glucose and fructose
Potential alcoholThe alcohol obtained if all the sugar in the must is fermented
Temperature correctionAdjusting a measurement to the standard temperature (20 °C)
Pearson squareA method for calculating the blending ratio of two solutions for a target concentration

Practical use of the calculations

It is the Rkatsiteli harvest on a warm September afternoon. A 200-berry sample taken in the vineyard reads 21.0 Brix on the refractometer, and the sample temperature is 28 °C. The calculator corrects for temperature: 8 degrees × 0.075 ≈ +0.6, so the real sugar content is 21.6 Brix — a difference that changes the style planned for the qvevri wine. 21.6 Brix corresponds to a specific gravity of 1.090, 90 Oechsle and about 212 g/L of sugar; with the European coefficient the potential alcohol is 12.6%, and allowing for real fermentation losses (17.5 g/L) it is 12.1%. The winemaker has two plots: 5000 litres of this must and 1500 litres picked later at 24.5 Brix. He wants a single must at 22.2 Brix. The blending mode calculates the mass balance and shows that about 1290 litres of the second must are needed, while adding all 1500 litres gives a blend of 22.3 Brix — so the whole second plot can be used without breaking the style. (For 22.5 Brix the 1500 litres would no longer be enough — about 2220 litres would be needed.) Finally, the sample-size mode confirms that 200 berries with $\sigma = 2$ give a precision of ±0.28 Brix — sufficient for this decision.

For beginners — from a refractometer reading to alcohol

1. Open the "Calculator", mode "Units and alcohol". Choose the instrument (refractometer or hydrometer), enter the reading and the sample temperature.

2. Press "Calculate". The cards show the corrected sugar content on every scale — Brix, specific gravity, Oechsle, Baumé — and the sugar concentration in grams per litre.

3. The lower card shows potential alcohol with two coefficients: the legal one (16.83 g/L) and one that allows for real fermentation losses (17.5 g/L by default). A real wine usually ends up between these two numbers.

4. If you want to combine the must of two plots or two qvevris, switch to the "Blending musts" mode and enter the sugar content and volume of both, and the desired sugar content.

5. The "Sample size" mode tells you how many berries to sample before you can trust the result. The rule is simple: four times as many berries doubles the precision.

For professionals — precision and sources of error

1. The temperature correction for the refractometer is linear (0.075 Brix/°C, the average of the ICUMSA tables in the 15–25 Brix range); for the hydrometer it is calculated from the temperature dependence of the density of water. Outside 10–30 °C both approximations deteriorate — it is better to bring the sample to 20 °C first.

2. The refractometer is calibrated on sucrose; the non-sugars of must (acids, minerals) raise the reading, so Brix slightly overestimates the real sugar content. The calculator's "non-sugar fraction" parameter (10% by default) corrects for this; raise it to 12–15% for high-acid or botrytis-affected grapes.

3. Once fermentation has started the refractometer is no longer usable: ethanol has a higher refractive index than sugar, so the reading greatly overestimates the real sugar. In fermenting must use only a hydrometer or laboratory analysis.

4. When blending, a mass balance is more accurate than a volume balance, because Brix is a mass quantity. For differences under 2 Brix the discrepancy is negligible; above 5 Brix it reaches several tens of litres in large volumes.

5. The standard deviation between berries depends on the plot: on a uniform plot it falls to 1.5 Brix, on an unevenly ripened one (sunny and shaded clusters, top and bottom of a slope) it exceeds 3 Brix. If you do not know σ, measure 20–30 individual berries once and enter the actual spread in the calculator.

1. Why is the refractometer unusable in fermenting must?

Ethanol has a higher refractive index than sugar, so after fermentation starts the refractometer shows much more sugar than is really there. A hydrometer is used in fermenting must.

2. A refractometer reads 21.0 Brix on a sample at 30 °C. Roughly what is the real sugar content?

A warm sample reads low; the correction is ≈ 0.075 × (30 − 20) ≈ +0.75 Brix.

3. Theoretically, how many grams of sugar per litre are needed for 1% vol of alcohol?

1% vol = 7.89 g/L of ethanol; 51.1% of the sugar becomes ethanol, so 7.89 / 0.511 ≈ 15.4 g/L. In practice 16.5–18 g/L is needed because of losses.

4. 50 berries are not enough for ±0.3 Brix. By what factor must the sample grow to halve the error?

The error is proportional to 1/√n, so to halve it the number of berries must be quadrupled.