Silicones in anhydrous products: what dimethicone does that an oil cannot

Dimethicone from 5 to 350 centistokes, phenyl trimethicone for gloss, the US skin protectant monograph at 1 to 30 percent and the EU D4, D5 and D6 restriction.

Dimethicone is the only material in a common balm that is simultaneously a cosmetic emollient, a recognised over the counter drug active in the United States, and, in one of its cyclic relatives, a substance the European Union is in the middle of restricting. Those three facts pull in different directions, and most of what is written about silicones argues one of them while ignoring the other two. This page covers the viscosity grades and what each does to a film, what the monograph allows you to claim, where a silicone genuinely beats an ester, and what will not stay mixed in a wax base.

• INCI: Dimethicone
• Surface tension 20-21 mN/m
• US monograph 1-30%
• EU leave-on D4/D5/D6 limit 0.1%

What a silicone is, and why it behaves unlike an oil

Dimethicone is a chain of alternating silicon and oxygen atoms with two methyl groups hanging off each silicon. Nothing in a plant oil looks like that. A triglyceride is a carbon backbone carrying polar ester groups and, in most balm oils, one or more double bonds; the silicone has no carbon backbone, no ester group and no unsaturation at all.

Three consequences follow, and they are the whole case for using it. The siloxane bond rotates far more freely than a carbon to carbon bond, so the chain drapes over a surface rather than balling up on it. The methyl groups face outward and present one of the lowest energy surfaces in cosmetic chemistry, roughly 20 to 21 mN/m of surface tension against 30 to 33 mN/m for a triglyceride oil. And with no bis-allylic hydrogens there is nothing for oxygen to attack, so a silicone does not go rancid, does not develop a cardboard note on the shelf, and does not need an antioxidant. It is also colourless, odourless and tasteless, which matters in a lip product where every other emollient contributes something.

The low surface tension is what a reader should hold on to. A liquid spreads spontaneously on a surface of higher energy than itself, so dimethicone spreads over skin, over wax, over lipstick and over the oils already on the lip, and then nothing much spreads over dimethicone. That single property produces the slip, the film continuity and the non transfer behaviour discussed below, and it is the reason a silicone can be felt at 2 percent when an ester needs 10.

Viscosity grades from 5 to 350 cSt

Dimethicone is sold by kinematic viscosity in centistokes, measured at 25 C by capillary viscometer under ASTM D445 or ISO 3104. One centistoke is one square millimetre per second; to convert to dynamic viscosity in mPa s, multiply by the density, about 0.96 to 0.97 g/mL. The grade number is the whole specification most suppliers give you, and it maps directly onto how the material behaves.

Read the table as a trade between spread and film. Low grades cover more area from the same weight and leave a thinner deposit that is quickly worn off by eating and talking. High grades deposit a thicker, more persistent layer at the cost of glide, and past about 1000 cSt they start to read as slippery in the way that people who dislike silicones actually mean. For a lip balm that has to feel light, 100 cSt at 3 to 5 percent is the safe starting point. For a barrier product meant to survive handwashing, 350 cSt at 8 to 12 percent does far more.

Occlusion: what the silicone film actually stops

Silicones are routinely called occlusive, and that is only half right. A dimethicone film is continuous and hydrophobic, but polydimethylsiloxane is unusually permeable to gases and to water vapour, which is why the same polymer is used for gas exchange membranes. It slows water loss without sealing the surface, and it is best described as semi occlusive.

The practical reading: if the job is to stop water leaving damaged skin, petrolatum wins and nothing in the silicone range comes close. If the job is to hold a film in place through friction, water contact and repeated handwashing without feeling greasy, the silicone wins because petrolatum transfers off onto everything it touches. Both claims are measurable, and the method behind the numbers matters more than the numbers do: see measuring occlusivity for how occlusion factor is determined and transepidermal water loss for what the reading means. Where each class of material sits is the subject of occlusive, emollient and humectant.

The US skin protectant monograph and the 1 to 30 percent window

In the United States, dimethicone is listed in 21 CFR part 347, the skin protectant monograph, as an active ingredient at 1 to 30 percent. That listing is not permission to add it to a cosmetic. It is the specification a product must meet if it wants to make the protectant claims the monograph allows, which are narrow and specific: helping prevent and temporarily protect chafed, chapped, cracked or windburned skin and lips, and for lip protectant products, relieving dryness and preventing chapping.

Cross that line and the product is an over the counter drug in the United States. It needs a Drug Facts panel with the active ingredient, purpose, uses, warnings and directions in the prescribed order, a drug listing and registered manufacturing establishment, and it falls outside the cosmetic regime entirely. The same tin with the same 5 percent dimethicone and no protectant claim is a cosmetic. The regulator is looking at the claim, not the ingredient. This distinction, and the wording that crosses it, is worked through in cosmetic versus drug claims.

There is no equivalent monograph in the EU or UK. Dimethicone is unrestricted under Regulation (EC) 1223/2009 and carries no annex entry, but a protectant claim still has to be substantiated against the common criteria in Regulation (EU) 655/2013, and the substantiation has to sit in your product information file. The route through that is in selling balms in the UK and EU.

Phenyl silicones for gloss and pigment wetting

Plain dimethicone has a refractive index near 1.40, lower than every common balm oil. Add enough of it and a glossy stick goes flatter, which is a genuine and frequently unexpected side effect. Putting a phenyl ring on the backbone raises the refractive index into a useful range: phenyl trimethicone sits near 1.46 and trimethylsiloxyphenyl dimethicone at 1.46 to 1.49, against roughly 1.478 for castor oil and 1.50 for a polybutene.

At 1 to 5 percent, a phenyl silicone does two jobs at once. It contributes gloss without the tack castor oil brings, and it wets pigment. Iron oxides and treated micas disperse more readily in a phenyl silicone than in a plain one, and the resulting dispersion resists the settling that leaves a tinted stick streaked. The optics behind all of this, including why refractive index alone does not make a balm shiny, are in lip colour and gloss physics, and the pigment side is covered in iron oxides and mineral pigments.

Phenyl silicones also act as a bridge. They are far more compatible with esters and triglycerides than plain dimethicone is, so 2 to 3 percent of phenyl trimethicone in a formula carrying 8 percent dimethicone will often hold the whole thing clear when the dimethicone alone would haze. If a tinted balm comes out dull rather than glossy, the phenyl grade is usually a better answer than more oil, and the diagnosis ladder is in matte instead of glossy balm.

Cyclopentasiloxane and the EU restriction

The cyclic siloxanes are a different material class from the linear polymer, and they are the ones under regulatory pressure. D4 (octamethylcyclotetrasiloxane), D5 (cyclopentasiloxane) and D6 (cyclohexasiloxane) are volatile rings that evaporate from skin over minutes, leaving whatever was dissolved in them behind. That is why they appear in colour cosmetics and non transfer products, where a carrier that vanishes is worth paying for.

Regulation (EU) 2018/35 introduced the rinse-off restriction on D4 and D5 through entry 70 of REACH Annex XVII. Regulation (EU) 2024/1328 extended that entry to leave-on cosmetic products and brought D6 in with them, at the same 0.1 percent by weight threshold, with 6 June 2026 as the date for leave-on cosmetics and later dates for other uses. The grounds were persistence and bioaccumulation, not skin safety. GB REACH is a separate legal system that does not automatically inherit EU restrictions, so a UK maker has to check the current GB text rather than assume alignment.

For anhydrous balms this is mostly academic and you should treat it as settled anyway. A volatile carrier has little to do in a product whose entire structure is wax and butter, since the thing it evaporates away from is already solid. If you are formulating a non transfer lip colour and reaching for cyclopentasiloxane, reformulate now onto isododecane or a fast ester rather than in 2026.

What will and will not stay mixed in a wax base

Dimethicone is not miscible with most balm materials in the way two oils are. Compatibility falls off with silicone viscosity, with the polarity of the base, and with total loading. As a working guide for an anhydrous formula built on triglycerides and wax: up to 5 percent of a 100 to 350 cSt fluid will usually disappear into the base; 5 to 12 percent needs testing; above 12 to 15 percent you are likely to see haze on cooling, beads on the surface after a warm cycle, or a slick that wipes off the top of a tin.

Polarity is the variable that catches people. Castor oil is the most polar liquid on a typical balm bench and the least compatible partner for a silicone, so a formula that leans on castor for gloss and dimethicone for slip is exactly the one that separates. Mineral oil, petrolatum, isododecane and the branched esters are far more tolerant. Alkyl-modified silicones such as cetyl dimethicone and stearyl dimethicone are built for this problem: their fatty chains anchor into the organic phase while the siloxane backbone does its usual work, and 1 to 3 percent will often hold a marginal formula together. The general framework is in solubility in anhydrous formulas.

Where a silicone beats an ester, and where it does not

The honest comparison is with the modern plant-derived esters, because that is the substitution most makers are actually considering. Coco-caprylate reproduces the fast, light break of a low viscosity silicone convincingly. Dicaprylyl carbonate and isoamyl laurate give slip that many testers cannot distinguish from dimethicone in a blind sensory panel. What none of them reproduce is persistence: an ester spreads, absorbs and is gone, while a dimethicone film is still there after two hours and after a hand wash. If your product is a hand balm for people who wash forty times a shift, an anti chafe stick for a marathon, or a lip colour that must not print on a cup, the silicone is doing something the ester cannot. If it is a face balm or a lip balm whose job is comfort, the ester is a straight swap. The full set of alternatives sits in cosmetic esters, and for gloss and cling specifically, polyisobutene and gloss polymers covers the other synthetic route.

Useful starting levels by product, all weight percent of the finished formula: lip balm 2 to 5 percent of a 100 cSt grade; tinted or long wear lip product 5 to 10 percent, with 2 to 3 percent of that as a phenyl grade; hand and barrier balm 5 to 12 percent of 350 cSt; anti chafe stick 8 to 15 percent, at the top of the range where friction is the whole point. Silicone contributes nothing to structure, so treat it as part of the liquid fraction when you set the wax level with the wax ratio calculator.

Biodegradability and the environmental question

Both sides of this argument are usually stated dishonestly, so here are the facts as they stand. Polydimethylsiloxane does not pass the standard OECD ready biodegradability tests, which are 28 day aqueous screens, and that failure is the basis for every "non biodegradable" claim made about it. It is a real result. What the claim leaves out is the degradation route that does operate: PDMS reaching a wastewater plant partitions almost entirely into sewage sludge rather than the effluent, and in soil it is hydrolysed on clay surfaces to dimethylsilanediol, which is water soluble, volatilises and is degraded further by soil microorganisms and by hydroxyl radicals in the atmosphere. Silicone does not persist in the way the phrase implies, but it also does not disappear on the timescale an aquatic biodegradation test measures.

The cyclic siloxanes are a separate and more serious case. D4, D5 and D6 were identified as substances of very high concern on persistence and bioaccumulation grounds, and that assessment is what drove the restriction described above. Applying that conclusion to linear dimethicone, as a good deal of consumer marketing does, is not supported.

Two further points that come up. Liquid dimethicone is not a microplastic under Regulation (EU) 2023/2055, which restricts solid synthetic polymer microparticles; cross-linked silicone elastomer powders are a different question and worth asking a supplier about directly. And silicone is not a petroleum product in the sense people mean when they say it: the silicon comes from quartz sand, though the methyl groups come from methanol that is usually fossil-derived. Whether that distinction matters is a values question rather than a technical one, and it is the subject of natural versus petroleum.

The decision rule

Use a silicone when the product has to stay put through friction or water and the alternatives have already failed a wear test. Do not use one because a formula feels slightly draggy, since that is usually an oil blend problem and a cheaper fix. Four situations rule it out regardless of performance: a certified natural line, because COSMOS and NATRUE both exclude silicones; a customer base that has decided against them, where the formulation argument is not the argument you are having; a product whose selling point is occlusion, where petrolatum is simply better; and a high gloss stick, where plain dimethicone works against you and only the phenyl grades help.

The safety position is settled and dull: dimethicone is poorly absorbed, has been reviewed repeatedly for cosmetic use, is not a sensitiser at these levels, and does not occlude pores in the way the folk claim suggests, as comedogenic ratings explains about occlusives generally. The arguments worth having about silicones are environmental and regulatory, not dermatological. A maker who can say what the ingredient does, and where the real objection to it lies, is in a better position than one repeating a slogan from either direction.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.