Silicone antifoam in beverage and food applications: the food sets the limit, not the drum

Your filler is foaming over on a new flavored drink, and the supplier’s sample arrives with a data sheet that says “food grade” and little else. That label answers none of the questions an export customer or an auditor will ask. May a silicone antifoam go into this drink at all? Up to what level, and measured how? The answer changes by product: the EU writes it food category by food category, and in the US, wine, meat and general food each follow a different federal text.

The additive is dimethyl polysiloxane (E 900 in the EU), sold to food plants as silicone antifoam food grade under CAS 63148-62-9. A purchase order may call it a food defoamer or an antifoam; the regulations care about three things neither name tells you: the food it goes into, the level and what that level measures.

Not every defoamer used for foam control in food processing is a silicone. 21 CFR 173.340(a)(1) (eCFR as of 2026-09-30) also admits substances generally recognized as safe in food for the use. Its (a)(2) table lists further components, among them white mineral oil, allowed only in wash water for sliced potatoes at no more than 0.008 percent of the wash water. This post follows the silicone, whose limits the EU writes food by food.

Which beverages and foods may a food defoamer go into, and at what level?

In the EU, the food-additive rows for E 900 sit in Annex II of Regulation (EC) No 1333/2008. The consolidated text of 18 August 2026 lists it in categories that include those below, each with a maximum in mg/kg or mg/l as appropriate, and Annex III adds a separate row for flavorings.

Of the thirteen Annex II rows above, twelve sit at 10 and chewing gum sits at 100. The numbers are not where buyers go wrong. The restrictions written inside the rows are.

In the 18 August 2026 text, the juice row names pineapple juice and two Nordic sweetened juice products, so an apple or orange juice line gets nothing from it. The fats row covers only oils and fats for frying, and the cider row excludes named traditional and ice ciders. The rows above say nothing about beer either way, and a brewer cannot borrow the cider row: find the legal basis your own use rests on before the first export lot ships.

The flavorings row is easy to misread. Annex III, in the same 18 August 2026 text, sets its maximum of 10 mg/kg in the flavoring itself, not in the finished drink, so ask a flavor supplier which figure its statement refers to.

What does each US limit measure?

In the US the question splits by product, and the three texts below put their numbers on different bases. The table follows the eCFR as of 2026-09-30.

For food in general, the FDA’s defoaming-agent section, 21 CFR 173.340 (eCFR as of 2026-09-30), sets its limit in the food as eaten, with exceptions for gelatin dessert mixes and cooking salt. Milk is the one product it names at zero. The 10 and 50 ppm rows in USDA’s table, 9 CFR 424.21 (eCFR as of 2026-09-30), belong to meat and poultry products, so a soup line cites that table only when the soup is one of them.

The two wine figures in the Alcohol and Tobacco Tax and Trade Bureau (TTB) list of materials for wine and juice treatment, 27 CFR 24.246 (eCFR as of 2026-09-30), are one limit written twice.

At 30% active, 60 mg/L of product carries 18 mg/L of active defoamer, the same as the 100% active line. A cellar that reads only the 60 mg/L line and doses a 100% active fluid at that rate adds more active defoamer than either line allows. Write the active basis next to every dose on the batch sheet.

The second wine number is written in silicon, the element, not in silicone. The E 900 specification in Commission Regulation (EU) No 231/2012 (consolidated text of 18 August 2026) puts total silicon at 37.3% to 38.5% of the polymer, so a silicon result and a silicone dose are different numbers for the same material. Ask your lab which one its report states before you compare it with the 10 ppm line.

Does fermentation tolerate a silicone antifoam?

27 CFR 24.246 (eCFR as of 2026-09-30) lists the wine use as “To control foaming, fermentation adjunct,” so the wine figures are fermentation doses too. A 2017 peer-reviewed study, Industrial antifoam agents impair ethanol fermentation and induce stress responses in yeast cells (Applied Microbiology and Biotechnology), tested this on yeast. Commercial antifoams including a 30% silicone emulsion caused no change in the growth rate of an industrial Saccharomyces cerevisiae strain up to 375 mg/L; the silicone emulsion lowered it at 750 mg/L.

The work used microplate cultures of one ethanol strain in a glucose medium, not a wine must or a wort, so treat it as a margin estimate, not a cellar rule. On growth rate, the TTB ceiling of 60 mg/L for a 30% active product sits well under the level where that strain slowed. Growth rate is one measure only: the paper’s own title reports impaired ethanol fermentation, so trial a new antifoam on your own must before a full tank.

What does “food grade” leave out of a silicone antifoam?

Two silicone antifoam compounds submitted as test materials in EFSA’s 2020 re-evaluation of E 900 show how much a compound can carry besides the fluid. One was 93% polydimethylsiloxane and 7% silica treated with polydimethylsiloxane; the other was 91% polydimethylsiloxane, 6% silica and 3% hydroxy-terminated polydimethylsiloxane.

Both carry silica alongside the silicone oil, so a composition statement should name both. A 2004 review of how oil-based antifoams destroy foam found that the entry barrier decides whether a pre-emulsified antifoam acts fast or slow, one reason to trial a new supplier’s product on the line rather than accept its data sheet.

The same EFSA record shows how far two compounds can sit apart: cyclosiloxane content was 9.75% in one and below 0.02% in the other. Both were documented to EFSA as silicone antifoam compounds, so the name alone does not tell you the cyclics content of the lot you buy.

Cyclosiloxanes are small silicone rings, among the first components to leave when the fluid is heated. The E 900 purity criteria in Commission Regulation (EU) No 231/2012 (consolidated text of 18 August 2026) cap loss on drying at 0.5% at 150 °C for 4 h, so ask for the cyclics figure and the loss-on-drying result on each lot and read them together.

The fluid specifications do not line up across texts either. The JECFA specification for polydimethylsiloxane (2008) allows 100 to 1,500 cSt at 25 °C, while 21 CFR 173.340 (eCFR as of 2026-09-30) asks for 300 to 1,050 centistokes at 25 °C and no more than 18 percent loss in weight after 4 hours at 200 °C.

The EU’s own E 900 specification in Commission Regulation (EU) No 231/2012 (consolidated text of 18 August 2026) sets only a floor, not less than 1.00 x 10^-4 m2/s at 25 °C. A fluid inside the JECFA window can sit outside the US one at either end, so a lot sold under more than one text needs its viscosity reported against each.

JECFA’s specification also describes the fluid, not your drum: it does not apply to aqueous formulations containing emulsifiers and preservatives, and it has silica gel removed by centrifugation before the identity and viscosity tests. A viscosity read on an emulsion, or on a compound still carrying its silica, is not that test.

What should the order for a silicone antifoam food grade lot say?

Write the order so that each line is answered by a document, not by the word on the label.

1. The application and the market. Name the EU food category or US text each use relies on, and have the supplier confirm in writing that its statement covers it.

2. The composition: percent polydimethylsiloxane and percent silica, any other silicone present and the cyclosiloxane content. For an emulsion, name each emulsifier, thickener and preservative, so each can be matched to the 21 CFR 173.340(a)(2) table (eCFR as of 2026-09-30), which caps formaldehyde as a preservative at 1.0 percent of the dimethylpolysiloxane content.

3. The fluid specification against each text you sell under: viscosity at 25 °C measured on the fluid, and the heat-loss result on the method each text uses.

4. Percent active and its basis, so a limit written for 100% or 30% active material converts to your dose without a guess. A silicone antifoam emulsion (10-50% active) and an undiluted compound need different dose figures for the same limit.

5. A certificate of analysis (CoA) for each lot. For EU sales, add the E 900 purity results that the 18 August 2026 text of Regulation (EU) No 231/2012 sets: arsenic not more than 3 mg/kg, lead not more than 1 mg/kg, mercury not more than 1 mg/kg and total silicon of 37.3% to 38.5%.

A silicone antifoam food grade order written this way can be checked line by line against the row or section each product relies on. One that says “food grade” and a percent active can be checked against nothing.

Methodology: EU figures are from Regulation (EC) No 1333/2008 (Annexes II and III) and Commission Regulation (EU) No 231/2012, consolidated texts of 18 August 2026; US figures are from 21 CFR 173.340, 27 CFR 24.246 and 9 CFR 424.21 in the eCFR as of 2026-09-30. Composition data are from EFSA’s 2020 re-evaluation of E 900 and yeast data from a 2017 peer-reviewed study. This is general information on these texts, not regulatory advice.