Winemaking Laboratory
Stabilisation and clarification / Protein stabilisation

Chapter IV · Stabilisation and clarification

Protein stabilisation

Why white wine turns hazy in the bottle in summer heat, how to check protein stability with a heat test, how to choose the minimum bentonite dose with a bench trial, and how much wine is lost in the lees.

Test result

Test: filtered wine is heated in a closed tube, then cooled to 4 °C (or room temperature) and the turbidity measured again. The wine is considered stable if turbidity rose by no more than 2 NTU (some winemakers use a stricter 1 NTU). When judging by eye, "faint opalescence" is ≈ 2–5 NTU.

Turbidity increase—NTU
Verdict—
Risk—
Next step—

Heat test for each dose

Bentonite dose (g/L) and turbidity increase (ΔNTU)

Dose (g/L)ΔNTU

In a bench trial different bentonite doses are added to several wine samples; after 1 day the samples are filtered and each is heat-tested. The turbidity increase falls exponentially with dose: ΔNTU ≈ a·e^(−d/d₀). The model fits this curve to your points and finds the minimum dose at which ΔNTU is below the threshold.

Minimum dose—g/L
Recommended dose—g/L
Turbidity without bentonite—NTU
Model fit—

Bentonite

Bentonite is first swollen in water (5–10%) for 12–24 hours, then added to the wine slowly, while stirring. Lees volume depends on the type: sodium ≈ 5%, intermediate ≈ 4%, calcium ≈ 2.5% of wine volume per 1 g/L of dose (approximately). The lees hold ≈ 70% wine; re-settling recovers half of it, filtration 80%.

Bentonite—g
Water for the slurry—L
Lees volume—L
Wine lost—L

Protein haze

A white wine that is perfectly clear in the cellar can turn hazy in the bottle in summer heat (in a hot car, a shop window, in the sun): first a faint opalescence appears, then a haze, finally white flakes. This is protein haze. It is harmless and barely changes the taste, but it spoils the wine's appearance and is unacceptable to buyers.

The haze is caused by the grape's defence proteins (PR proteins, "pathogenesis-related"), mainly thaumatin-like proteins and chitinases, which the vine accumulates against fungi and insects. These proteins are small (20–30 kilodaltons) and resist an acidic environment and proteases, so they pass through fermentation unchanged and remain in wine at 10–300 mg/L. In heat they denature: the molecule unfolds, its hydrophobic parts come to the surface, and the proteins stick to each other and to phenolics and sulphate. This is how visible particles form.

The amount of protein depends on the variety, ripeness, vintage and the condition of the vineyard: fungal disease and mechanical damage increase the amount of PR proteins. Total protein content is not a good measure of stability, because the proteins differ in sensitivity, so stability is checked by a direct test.

Protein haze is rare in red and amber wine: during skin contact tannins bind the proteins, which go into the lees during fermentation and maceration. That is why amber wine made on the skins in a qvevri usually needs no bentonite. Protein stabilisation is mainly an issue for white and rosé wine made by the European method.

The heat test

Stability is checked with a heat test: a filtered wine sample is heated in a closed tube, cooled, and its turbidity before and after heating is compared. Turbidity is measured with a turbidimeter in NTU (nephelometric turbidity units): a clean white wine reads 0.5–2 NTU, and a haze visible to the eye starts from ≈ 2–5 NTU.

TestConditionsNotes
Standard80 °C, 2 hours, then 4 °C or 20 °Cthe most common; the wine is stable if ΔNTU ≤ 2
Strict80 °C, 6 hours"catches" more protein; for long storage
Mild40 °C, 24 hours or morecloser to real hot storage; less sensitive

The test is generally harsher than real storage, and this is desirable: if a wine survives 80 °C, it will survive a summer car. Some laboratories also use trichloroacetic acid and tannin tests, but these often indicate too much bentonite.

Bentonite

Bentonite is a clay of volcanic origin, mainly montmorillonite. Its crystal consists of thin, negatively charged plates, between which water and ions penetrate, so bentonite swells in water and creates an enormous surface (≈ 700 m² per gram). At wine pH (3–3.6) proteins are positively charged and stick electrostatically to the bentonite plates. The protein-laden bentonite settles, and the proteins leave the wine.

BentoniteSwellingEffectivenessLees
Sodiumstronghighloose, large volume
Sodium-calciummediummediummedium
Calciumweaklowcompact

Bentonite's effectiveness also depends on pH: at low pH proteins are more strongly charged and bind better, so an acidic wine needs less bentonite.

Bentonite is not selective: besides proteins it also removes some aroma compounds, anthocyanins and body. So the dose should be the minimum that makes the wine stable, which is found with a bench trial. Typical doses are 0.2–1.0 g/L; doses above 1 g/L noticeably harm aroma. The turbidity increase falls roughly exponentially with dose:

$$\Delta NTU(d) \approx a\cdot e^{-d/d_0}, \qquad d_{\min} = d_0\cdot\ln\frac{a}{\Delta NTU_{\text{thr}}}$$

Lees and losses

Bentonite lees are loose and hold a lot of wine: each 1 g/L of sodium bentonite gives lees of ≈ 5% of the wine volume, calcium bentonite ≈ 2.5%. The lees are ≈ 70% wine. If you simply pour the lees away, at 1 g/L you will lose ≈ 3.5 litres per 100 litres. Re-settling the lees in a separate vessel or filtering them recovers part of the wine.

Bentonite can be added at different times: to the must before or during fermentation (the lees settle together with the yeast lees and losses are smaller, but choosing the dose accurately is difficult), or to the finished wine after a bench trial, the most accurate approach. Before adding bentonite the wine should be clear and fermentation finished; protein stabilisation is done before cold (tartrate) stabilisation.

Alternatives to bentonite: enzymatic (protease) treatment combined with heat (briefly heating the must), chitosan, carrageenan, ultrafiltration. At home bentonite is the simplest and cheapest.

How to use this page

1. In the "Heat test" mode choose the test conditions and how you judge turbidity: with a turbidimeter (NTU) or by eye. Enter the turbidity before and after heating: you will see the increase in turbidity, a verdict, the risk of haze in the bottle and the next step.

2. In the "Bentonite bench trial" mode enter each dose and its heat test result (ΔNTU). The model fits a curve to the points and shows the minimum and recommended dose. The recommended dose is passed automatically to the third mode.

3. In the "Dose and losses" mode choose the bentonite type, dose, wine volume and how you will handle the lees: you get the grams of bentonite, the water for the slurry, the lees volume and how much wine you will lose.

The heat test at home

1. Take two wine samples in clear glass tubes or small bottles that can be closed. Filter the wine first through paper (a coffee filter will do): the sample must be perfectly clear.

2. One sample is the control. Put the other in an 80 °C water bath (a saucepan, with a thermometer) for 2 hours. The tube must be tightly closed so the wine does not evaporate.

3. Cool the heated sample in the fridge to 4 °C for a few hours, then bring it to room temperature.

4. In a dark room, light both samples from the side with a narrow torch beam. If the beam shows more clearly in the heated sample (opalescence), or haze or flakes have appeared, the wine is protein unstable.

Bench trial

1. Make a 5% bentonite slurry: 5 g of bentonite in 100 mL of hot water, stirring; leave for 12–24 hours to swell. Shake well before use.

2. Pour 100 mL of wine into 5–6 identical bottles. Each 1 mL of 5% slurry in 100 mL of wine = 0.5 g/L bentonite. So for 0.2 g/L add 0.4 mL, for 0.4 g/L 0.8 mL, for 0.6 g/L 1.2 mL, and so on. Add nothing to one bottle.

3. Shake, leave for 1 day, then filter each and heat-test it.

4. Enter the results on this page. When judging by eye, enter an approximate NTU: perfectly clear ≈ 0, faint opalescence ≈ 3, clear haze ≈ 10, cloudy ≈ 30.

Practical tips

• Always swell bentonite in water first: dry powder clumps in wine and does not work.

• Add the slurry to the wine slowly, stirring constantly, then stir again several times over 1–2 days. Bentonite settles in wine within 1–2 weeks.

• Rack the wine carefully off the lees. Pour the lees into a separate vessel and let them settle for a few more days: you can draw off extra wine from the top.

• Choose the bentonite dose with a bench trial and do not "round up" much: excess bentonite strips the wine of aroma and body.

• Bentonite is sold in winemaking supply shops. Do not use pharmaceutical or cosmetic "clay".

1. Why does protein haze appear in a white wine in the bottle if it gets hot?

Grape PR proteins (thaumatin-like proteins, chitinases) survive fermentation and an acidic environment, but in heat they denature and clump into visible particles.

2. Why do red and amber wines usually not need bentonite?

During skin contact tannins bind unstable proteins and these go into the lees. That is why amber qvevri wine is usually protein stable.

3. In a bench trial ΔNTU = 3 at 0.4 g/L and 1.2 at 0.6 g/L. The threshold is 2 NTU. Which dose is sensible?

The minimum dose lies between 0.4 and 0.6; with a small margin ≈ 0.55–0.6 g/L is enough. Excess bentonite strips aroma and body, and you lose more wine in the lees.