Winemaking guide · Protein stability
How much bentonite? Protein stability in white and rosé wine
Most white and rosé wines need 10 to 100 g/hL of bentonite, or 0.1 to 1 g/L, and some need none at all. Only a bench trial tells you how much bentonite to add to wine like yours: a few samples at rising doses, a heat test at 80 °C (176 °F), and the lowest dose that stays clear.
Protein haze in wine: from invisible protein to white cloud
Every white wine holds proteins from the grapes. The ones that cause haze belong mostly to two families of defense proteins the vine makes against disease (PR proteins): thaumatin-like proteins (TLPs) and chitinases. They survive fermentation and the wine’s acidity, and you can’t see them.
The trouble starts with heat. The folded protein opens up, its sticky parts are exposed, and it clings to other proteins and to compounds such as phenolics and sulfate. The clumps grow until they scatter light: the wine looks hazy, and sometimes white flakes appear.
- Folded protein In a cool cellar the protein is compact and soluble. You can’t see it, and the wine is clear.
- Heat opens it up In heat the protein unfolds, and its sticky parts (yellow) are exposed. The wine is still clear.
- It sticks to others The open proteins bind to each other, to phenolics (green) and to sulfate in the wine.
- Visible haze As it cools, the clumps grow and scatter light. The wine is hazy, sometimes with white flakes.
This haze can form at any temperature, but it forms faster in heat (AWRI). Above about 65 °C (149 °F) the proteins unfold, so the stability test heats the wine to fast-forward months of storage in bottle.
Which wines get it
The proteins come from the grapes and are already in the juice. In red wine protein haze is rare: tannins bind to the proteins and settle out with them, so almost no unstable protein is left. The exceptions are low-tannin reds, reds from very cool regions, and sometimes Pinot Noir.
In sparkling base wine, the second fermentation in bottle adds alcohol, so some producers add, for example, about 1–1.5% alcohol to the sample before the heat test.
A hot summer is a heat test you didn’t ask for
In a UC Davis trial, eight white wines that had not been fined with bentonite were put into storage. After a month at 35 °C (95 °F), as in a truck or a warehouse in summer, 7 of them turned hazy. Even in closed cases in controlled storage at 13–17 °C (55–63 °F), 6 did. The wines fined with bentonite stayed clear, except one Riesling.
Fruit from warm regions, such as Israel, also often carries more protein (Virginia Tech). A white wine that leaves the winery without a stability test is a bet on the trip to the shop.
What sets the protein level in wine
Wine holds between 10 and 300 mg of protein per liter (Purdue). These are the main factors:
-
Variety Gewürztraminer is known for being rich in protein. Sauvignon Blanc and Pinot Gris often need a lot of bentonite, but any wine can surprise you.
-
Ripeness Protein rises as the grapes ripen, and quickly after véraison, when the berries start to soften and change color.
-
Skin contact and pressing Most of the protein comes out in the first 10 hours of skin contact. Whole-cluster pressing gives juice with less protein.
-
Machine harvesting and transport In an Australian study, machine harvesting with long transport raised the PR proteins. A study from New Zealand found the opposite.
-
Disease Powdery mildew raises the PR proteins. Botrytis, on the other hand, lowers them, because the fungus breaks proteins down.
-
Wine chemistry pH, alcohol, phenolics and sulfate all affect haze. In the heat test a high-pH wine hazes more, but in normal storage it is low pH that raises the risk (AWRI). Planning an acid adjustment? Do it before the trial.
The fix: bentonite fining
Fining means adding something that binds a compound in the wine and settles out with it. Bentonite is a clay built from very thin layers of an aluminum silicate mineral. The layers carry a negative charge, and positive ions such as sodium or calcium sit between them. In water the layers move apart (swelling), and a huge surface area opens up.
At wine pH most proteins carry a positive charge. When they meet bentonite, they take the place of the sodium ions and bind to the layers (ion exchange). The bentonite settles to the bottom with the proteins on it, and you rack the wine off the lees.
- Dry powder Thin clay layers, packed tightly together. Sodium ions (yellow) are trapped between them.
- Swelling in hot water Water (blue) gets between the layers. They move apart, and their negatively charged surfaces are exposed.
- The proteins bind Positively charged proteins cling to the surfaces and take the place of the sodium.
- It settles as lees The loaded clay is heavy and sinks. The wine above it is stable, and you rack it off the lees.
Sodium or calcium bentonite
The bag usually says which ion sits between the layers. That decides how much the bentonite swells, how much protein it takes up and how much wine you lose in the lees.
| Sodium bentonite (Na) | Calcium bentonite (Ca) | |
|---|---|---|
| Swelling in water | High | Low |
| Protein binding | Strong | Weaker, often needs more powder |
| Lees | Fluffy and loose, more wine lost when racking | Compact, less wine lost |
| Clarification | Good | Better |
| Watch out | A heavy treatment can double the wine’s natural sodium, which is low to begin with | Not recommended above pH 3.4. Adds calcium and raises the risk of calcium tartrate crystals |
There is also sodium-activated calcium bentonite, which falls in between. Whichever type you choose, run the trial with exactly the same bentonite, from the same lot, that you will use in the tank. MoreWine notes that one product may need four times as much as another.
Bentonite dosage: the usual range, and why only a trial decides
According to AWRI, trials usually test 0.1 to 1 g/L, or 10 to 100 g/hL. Aromatic, protein-rich wines such as Sauvignon Blanc can need 1–3 g/L (Enartis).
But the dose doesn’t follow from the amount of protein. Here are seven wines tested at AWRI:
Home winemaking forums sometimes say to add a gram per liter and be done with it. For the Chardonnay in the chart that is bentonite it never needed, and for the Riesling five times too much. Extra bentonite strips aroma and flavor, and makes more lees, which means more wine you throw away.
Heat stability test: how to check for protein haze
The method most winemakers use today is AWRI’s:
- Filter through 0.45 µm, as before bottling
- Measure the starting haze in NTU
- Heat in a water bath 80 °C · 2 hours
- Cool in water 20 °C · 3 hours
- Measure again the change from the start Δ < 2 NTU
NTU is a unit of turbidity. You measure it with a nephelometer (turbidity meter), which checks how much light the sample scatters. Some labs are stricter and ask for a rise of less than 0.5 NTU, but AWRI sees 2.0 as a more reasonable threshold.
Fast cooling makes more haze, so stick to 20 °C (68 °F). If you don’t have a bath at that temperature, one hour in the fridge and two hours at room temperature will do. The old method, 6 hours of heating and a night at 4 °C (39 °F), may overstate the dose (AWRI).
Without a nephelometer
Many wineries check by eye: in a 1998 survey in Australia, 29 of 34 wineries relied on a visual check. To make it work, use identical sealed tubes, an unheated control sample next to them, a dark room and a strong flashlight shining from the side. A rule of thumb: if you can see haze after heating, the wine is not stable.
A tip for home winemakers: a sous-vide cooker in a pot of water holds a steady 80 °C (176 °F) for two hours. Make sure your model reaches that temperature, and seal the tubes well.
Worked example: a 100 mL trial, a 1,500 L tank
A Sauvignon Blanc after fermentation, 1,500 L (about 400 US gallons) in the tank. Make a 5% bentonite slurry (bentonite mixed into water): 5 g of bentonite, made up to 100 mL with water. Take five 100 mL samples of the wine: one is the control, and the other four get slurry at rising doses. Let them settle, filter them and run the heat test. Tap each sample:
Dose in g/hL · slurry added to 100 mL · NTU rise after heating
1.2 mL of slurry in 100 mL of wine is 60 g/hL. A rise of 1.1 NTU: stable. For a 1,500 L tank: 900 g of bentonite. This is the lowest dose that passed.
Control sample, no bentonite: haze rose by 14.2 NTU. The wine is not stable.0.4 mL of slurry in 100 mL of wine is 20 g/hL. A rise of 6.8 NTU: not stable. For a 1,500 L tank: 300 g of bentonite.0.8 mL of slurry in 100 mL of wine is 40 g/hL. A rise of 2.9 NTU: not stable. For a 1,500 L tank: 600 g of bentonite.1.2 mL of slurry in 100 mL of wine is 60 g/hL. A rise of 1.1 NTU: stable. For a 1,500 L tank: 900 g of bentonite. This is the lowest dose that passed.1.6 mL of slurry in 100 mL of wine is 80 g/hL. A rise of 0.4 NTU: stable. For a 1,500 L tank: 1,200 g of bentonite. It passes, but it is more than you need.Stability threshold: a rise of less than 2.0 NTU after heating
- Sauvignon Blanc
- Volume 1,500 L
- 100 mL samples
- 5% slurry
- Doses 0–80 g/hL
- Slurry per sample: 1 mL of a 5% slurry in 100 mL of wine is 500 mg/L, or 50 g/hL. So 0.4, 0.8, 1.2 and 1.6 mL give 20, 40, 60 and 80.
- The results: 60 g/hL is the lowest dose that stays under 2.0 NTU.
- Bentonite in the sample:
5 × 1.2 ÷ 100 = 0.06 g - The dose in mg/L:
0.06 × 1,000,000 ÷ 100 = 600
which is 60 g/hL. - Bentonite for the tank:
60 × 1,500 ÷ 100 = 900 g - Or in the same 5% slurry (1.2 mL per 100 mL of wine):
1.2 × 1,500 ÷ 100 = 18 L
Add 900 g of bentonite
18 L of slurry is 1.2% of the tank volume, and that is a lot of water. The usual rule (Purdue’s, for one) is to add no more than 1% water to the wine, so for the tank make a thicker slurry: about 1 kg per 8 L of water, which is 900 g in about 7.2 L, under half a percent. Keep the water temperature and the swelling time the same as in the trial.
If you are torn between two doses, AWRI recommends the higher one. Here 60 passed with a comfortable margin (1.1 NTU), so there is no reason to go up to 80.
Work it out for your wine. Pick Bentonite as the Product, then enter the Stock solution strength, the Stock added to the sample that passed, the Sample size and the Volume to treat. For this example that is 5%, 1.2 mL, 100 mL and 1,500 L, and it gives 900 g, or 18 L of the same slurry.
If you already know the dose in g/hL, the additive dose calculator turns it into grams for any volume.
In the tank: prepare, add, settle, rack
- Weigh the bentonite to the calculation, from the same lot you tested in the trial.
- Make a slurry. Use hot water, about 60 °C (140 °F), with no chlorine. Sprinkle the powder into the water while stirring hard, not the other way round, so no lumps form. Don’t swell it in wine: there the bentonite never swells all the way.
- Let it swell for at least 4 hours, usually overnight. Don’t keep a ready slurry for more than 24 hours. Some products tell you to swell them in cold water.
- Add it slowly and stir hard for at least 10–15 minutes. The wine should be above 10 °C (50 °F). Stirring with an inert gas, such as nitrogen, cuts down oxidation.
- Let it settle. In a tank this takes about a week, depending on the height of the tank.
- Rack off the lees. Tip: let the lees settle again in a small vessel, or filter them, to save more wine. Taste that wine before you return it to the tank.
- Test again after every addition and blend. Copper, tannin or blending with another wine can upset a wine that was already stable. Lysozyme, used to block malolactic fermentation, can also haze a white wine, and it is best to leave two weeks between it and the bentonite.
In the juice or in the wine?
You can add bentonite to the juice, before or during fermentation. The upside: much of the aroma is still bound and doesn’t get stripped, and there is less lees to deal with after fermentation. The downside: juice holds more protein, so the bentonite works less well. It also takes up nitrogen and other compounds the yeast needs, and can slow the fermentation or raise H₂S. If you add it to the juice, top up the yeast nutrients.
Even after treating the juice, test the finished wine, and add more if it needs it.
Heat stabilization first, cold stabilization second
- Heat stabilization Bentonite against protein haze
- Cold stabilization Against tartrate crystals (KHT)
- Filtering and bottling After a final check
There are three reasons for this order. Calcium bentonite adds calcium to the wine and raises the risk of calcium tartrate crystals, so test tartrate stability only after fining. Tartrate stabilizers such as CMC, metatartaric acid and polyaspartate can cloud a wine that isn’t protein stable. And in the cold, the tartrate crystals (KHT) pack down the bentonite lees, so you lose less wine when racking.
Tartrate dropping out in the cold also shifts the pH, so Purdue recommends chilling the trial samples too if you plan to cold stabilize. The next step is covered in crystals in wine and cold stabilization.
Work out the sulfite addition before bottling with the guide How much sulfite to add to wine.
Alternatives to bentonite
- Enzyme and flash heating A protein-breaking enzyme (Aspergillopepsin I), together with about a minute of heating at 60–75 °C (140–167 °F) that opens the proteins up for it. The OIV approved the method in 2021 for juice and wine. It needs flash-heating equipment, and check what your country allows.
- Carrageenan A polysaccharide from red seaweed. 125–250 mg/L heat-stabilizes a wine. According to Wineland (2024), it is allowed only in Australia, and the OIV and the EU have not approved it. Check what your country allows.
- Ultrafiltration A membrane with 10 kDa pores holds back up to 90% of the proteins. It is expensive, and takes aroma too.
- Aging on lees (sur lie) Mannoproteins from the yeast cell walls give some protection against haze, and can cut the bentonite you need by up to half. They don’t replace it.
Common mistakes
- A trial with a different bentonite Another product or another lot gives a different dose.
- mL instead of grams If you made a thicker slurry for the tank, work from grams of bentonite, not from the mL of slurry in the trial.
- Testing without a control Without an unheated sample there is nothing to compare against.
- Unfiltered samples Before heating, the samples should read under 2 NTU, or leftover bentonite will throw off the reading.
Frequently asked questions
What about ready-made test kits?
According to AWRI, Proteotest and Prostab predict a dose close to the 2-hour heat test. Bentotest predicts higher doses, so it can lead to overfining.
How long should the trial samples sit before the heat test?
AWRI gives one to two hours, or overnight, until the wine starts to clear and lees build up at the bottom. Scott Labs recommends at least 24 hours.
What is the difference between g/hL and g/L?
A hectoliter is 100 L. 1 g/L is 100 g/hL, and 1 g/hL is 10 mg/L (ppm). In US gallons, 1 g/hL is about 0.04 g per gallon, so 60 g/hL is about 2.3 g per gallon.
Glossary
- Bentonite
- A clay that swells in water and binds proteins.
- Slurry
- Bentonite powder dispersed and swollen in water, before it goes into the wine.
- PR proteins
- The vine’s defense proteins, including TLPs and chitinases. They cause most protein haze.
- NTU
- A unit of turbidity, measured with a nephelometer (turbidity meter).
- Racking
- Moving the clear wine off the lees into another vessel.
- Cold stabilization
- Chilling the wine so tartrate crystals (KHT) drop out before bottling.
Calculators for this guide
Sources
- Conducting a bentonite fining trial — AWRI: 5% slurry, 1 mL in 100 mL = 500 mg/L, a 0.1–1 g/L range, the 2.0 vs 0.5 NTU threshold.
- Protein stability fact sheet — AWRI: the heat test, cooling, choosing between two doses, test kits, retesting after additions.
- AWRI Technical Review 235: the stages of haze formation, unfolding above 65 °C, the effect of pH, the protein and dose table.
- Fining with bentonite — Purdue Extension (Butzke): sodium vs calcium, making the slurry, 1% water, stirring, settling, Gewürztraminer, red wines, chilling the samples.
- Wine proteins and protein stability — Virginia Tech Enology: ripeness, warm regions, skin contact, 5–10% lees, bentonite during fermentation, sparkling wine, the 2-hour test.
- Stability tests and haze formation — UC Davis: storage at 35 °C, the visual-check survey.
- Wine proteins and haze — IntechOpen: PR proteins, heat unfolding, disease, machine harvesting, carrageenan, ultrafiltration, mannoproteins.
- Stabilizing proteins from the beginning — Enartis: comparing bentonite types, 1–3 g/L in aromatic varieties, 3–5% wine loss, treating the juice.
- Heat stability in white wine — MoreWine: sodium vs calcium lees, differences between products, aging on lees.
- Inobent technical data sheet — Scott Laboratories: keeping the slurry, swelling in cold water, settling time in the trial.
- OIV-OENO 541B-2021: Aspergillopepsin I with flash heating.
- OIV-OENO 541A-2021: Aspergillopepsin I with flash heating.
- The use of carrageenan — Wineland: the legal status of carrageenan (November 2024).
- Wine tartrate stability — Enartis: protein stability before tartrate stabilizers.
- Lysozyme to arrest MLF — WineMaker: lysozyme can cause protein haze in white wine, and two weeks between it and bentonite.
- A guide to making dry white wine (in Hebrew) — Israeli home winemaking forum: heat stabilization before cold, and lees that pack down in the cold.