Why castor oil makes your lip balm sticky

Why castor oil makes your lip balm sticky

You made a lip balm to a recipe you found, with castor oil in it because every lipstick recipe has castor oil in it. It goes on well, the colour is even, and then your lips stick to each other. Not a disaster — a tug, a tackiness, a film that is still there half an hour later and picks up every stray hair on a windy day.

So you go looking for the culprit, and everything points at the castor. And it is the castor. But before you take it out, it is worth knowing that in a lipstick that tack is not a defect. It is the job you hired it for, and if you remove it without replacing what it was doing, the next batch will have a different fault.

Here is where the stickiness comes from, what it buys you, and when to cut it.

One hydroxyl group is the whole story

Every other article in this series sorts oils by a fatty acid that several plants share. Castor is the exception — the single-member group in our guide to plant oils by fatty-acid group — and the reason is one extra chemical group.

Its dominant fatty acid, ricinoleic acid, is 12-hydroxy-cis-9-octadecenoic acid. In plain terms: an eighteen-carbon chain with one double bond at position 9, exactly like oleic acid, plus a hydroxyl group — an –OH, the same group that makes alcohols behave the way they do — hanging off carbon 12. Nothing else in an ordinary oil cupboard carries one.

How much of it there is turns the exception into a rule. Patel and colleagues in Lipid Insights (2016) give castor oil as up to 90 per cent ricinoleic acid, 4 per cent linoleic, 3 per cent oleic, 1 per cent stearic and under 1 per cent linolenic. An independent measurement by gas chromatography and mass spectrometry from Ciastowicz and colleagues in Materials (2025) — a study characterising oils for polyurethane adhesives, not for skin, which is worth remembering every time we quote one of its numbers — is close: ricinoleic acid, reported as C18:1 OH, at 84.76 per cent, with oleic at 5.68 and linoleic at 5.94. In their five-oil comparison it appeared in castor and in none of the others.

Notice what that does to the arithmetic. In a balanced oil, the behaviour of the liquid is an average over several fatty acids, and no single one dominates. Castor is not an average of anything. Between eight and nine chains in ten carry the hydroxyl group — so the group is not a feature of the liquid, it is the liquid.

A housekeeping note, since readers of this series reach for the Codex tables: castor has no entry in the Codex Standard for Named Vegetable Oils, and should not, because that standard covers oils presented for human consumption. Ogunniyi opens a review in Bioresource Technology (2006) on exactly that note — even though castor oil is inedible, it has long been an article of commerce, in large measure because of its versatility. Your composition references come from the literature and from the supplier's own specification, and nowhere else.

At the bench: when you substitute castor in a formula, you are not swapping one oleic oil for another. You are removing a functional group that nothing else in your cupboard has, and you should expect several properties to move at once.

Why it pours like syrup

Viscosity is the property everyone notices first, and it only means something measured against comparators in the same laboratory on the same day.

Ciastowicz and colleagues did that, on a cone-plate rheometer to EN ISO 3219-2 at 20 °C:

Hydroxyl value is simply the laboratory count of how many hydroxyl groups a material carries, expressed as milligrams of potassium hydroxide per gram. Castor's is about two orders of magnitude above the others. Its viscosity is about twelve times the highest of them.

Because one laboratory can be an outlier, it is worth checking the ordinary band somewhere else. Yalçın, Toker and Doğan in the Journal of Oleo Science (2012) measured six vegetable oils on a rheometer at 25 °C and found them spanning barely thirteen units: olive 61.2, hazelnut 59.7, cottonseed 57.3, canola 53.5, soybean 48.7, sunflower 48.2 mPa·s — with viscosity correlating with monounsaturated content (R = 0.89) and inversely with polyunsaturated content (R = −0.97). Those oils differ substantially in profile and still sit inside a narrow band. The two studies used different temperatures and are not comparable line by line, but the point stands: ordinary plant oils live between roughly 50 and 110 mPa·s, and castor is somewhere else entirely.

What puts it there is not chain length and not unsaturation — it shares both with the oleic oils. The castor sample combined a much higher measured viscosity with a much higher hydroxyl value than its comparators, and their infrared spectra show the group directly, as a broad absorption band around 3380 to 3400 cm⁻¹ which the authors identify with hydroxyl groups and describe as particularly pronounced in castor oil.

The mechanism joining those observations is structural reasoning, and we label it as such rather than dress it up as a citation. A hydroxyl group both donates and accepts hydrogen bonds — weak attachments between molecules that behave like temporary stitching. In a triglyceride carrying one on most of its chains, molecules associate through those bonds, and the stitching has to be undone before layers of liquid can slide past each other. That is a plausible explanation of both the viscosity and the tack; the study reports a band and a number, not a mechanism.

The closest measured support we could open comes from a different system. Zhang and Weiss in ChemPhysChem (2016) studied ricinelaidic acid — the trans isomer of ricinoleic acid — and its ammonium salts, using infrared spectroscopy and powder X-ray diffraction to show that structural changes produced different packing arrangements and different strengths of hydrogen-bonding interactions within the assemblies. That is the free hydroxy acid and its salts rather than the oil, and assemblies rather than bulk flow. It shows this hydroxyl group forms hydrogen bonds whose strength depends on structure. It does not show that they set the viscosity of the oil.

And one consequence for the shelf, because it cuts against the folklore. Patel and colleagues note that the hydroxyl functionality makes castor a natural polyol and, in their words, provides oxidative stability and a relatively high shelf life compared with other oils by preventing peroxide formation. But in Ciastowicz's table the castor sample had the lowest iodine value of the five oils, 85.5 g I₂/100 g — and a peroxide value of 16.75 meq O₂/kg, higher than rapeseed (1.85), sunflower (5.18) or linseed (7.65) in that same table. Low unsaturation is not a shelf-life guarantee, and the numbers that measure each are unpicked in oxidative stability of oils.

At the bench: record the temperature beside every viscosity figure you keep, and check the peroxide value on castor deliveries like any other oil. Its reputation for indestructibility is not supported by the one comparative table we could open.

The tack is the function — and it has been measured

Sensory vocabulary treats tack as a fault. In colour cosmetics it is the reason the product works, and Maktabi, Liberatore and Baki in the International Journal of Cosmetic Science (2021) put numbers on it, which is what lets this section be more than an opinion.

They dispersed Red 7 Lake — a standard lipstick pigment — at 20, 30 and 40 per cent solids in castor oil, octyldodecanol and meadowfoam seed oil, measured firmness and stickiness on a texture analyser alongside viscosity, wetting and oil absorption, then made lipsticks from four of the dispersions.

Those are figures for one pigment, one loading, one laboratory — not a specification for anyone else's pigment. Stickiness is reported by the instrument as a negative, tensile force; the signs are theirs. Read across the row and the castor dispersion was roughly ten times firmer and ten times stickier than the other two at the same pigment loading.

Two of the other columns explain why formulators keep coming back to it. Red 7 Lake absorbed 1.54 g of castor oil per gram of pigment against 1.11 and 1.05 for the others — which the authors link to higher stick hardness — and all three agents wet the pigment well. An agent that wets pigment, is absorbed by it and holds the paste together is doing exactly what a lipstick needs. The stickiness reading is that same physics seen from the other side. You cannot have one without the other.

Reported practice matches the physics. The Cosmetic Ingredient Review Expert Panel — a United States industry expert panel, not a regulator and not a UK authority — recorded in its 2007 safety assessment that castor oil and hydrogenated castor oil were in use in 769 and 202 cosmetic products respectively in 2002, and that the highest reported use concentration for castor oil, 81 per cent, was associated with lipstick. That is a survey of what companies reported in one year. It is not a permitted maximum; in Great Britain the level is whatever your own product's safety assessment supports.

That panel concluded that castor oil, ricinoleic acid and the listed salts and esters are safe in the practices of use and concentrations described in its assessment. Two further facts from it are worth carrying away, because both answer questions people actually ask about this oil. Castor oil does not contain ricin, because ricin does not partition into the oil — the toxin stays with the pressed seed meal. And castor oil and ricinoleic acid can enhance the transdermal penetration of other chemicals, which is not a reason to avoid the oil but is a reason to look at what else is in the formula beside it.

Mascara is the other place tack is wanted, and here the published record thins out. In our bench work the castor level is a variable worth testing for adhesion, setting and transfer — but we found no published mascara measurement supporting a general effect, so treat that as a design for a paired trial rather than a finding.

At the bench: if the brief is a long-wearing, high-pigment lip product, do not fight the tack — budget for it, and tune it with the level rather than by removing the oil. If the brief is a light, slippy balm, castor is the wrong tool and no amount of reformulating around it will hide that.

When you do want less of it: change the wax, not the castor

The commonest mistake we see here is a substitution loop. The balm is sticky, so you cut the castor; the stick goes soft, so you add castor back to firm it up; the stickiness returns. That loop runs on a false premise — that castor level sets the firmness.

Budai, Antal, Klebovich and Budai in the Journal of Cosmetic Science (2012) examined sunflower, castor, jojoba and coconut oils with beeswax, candelilla and carnauba in stick bases, characterising hardness by compression strength alongside softening point, drop point and differential scanning calorimetry (a technique that measures the heat taken up as a material melts). Their conclusion names the levers: coconut oil, jojoba oil and carnauba wax have the greatest influence on the thermal parameters of stick bases. Castor was not among them.

So firmness lives mostly in the wax and the solid fat, and it is also entangled with pigment loading — Maktabi's own result was that the pigment's absorption of the dispersing oil contributed to stick hardness.

At the bench: when a stick is both too sticky and too soft, treat them as two separate jobs. Reduce the castor to fix the tack, then rebuild the firmness with the wax or the solid fat — and change one at a time, at fixed temperature, or you will not know which move did what.

Ethanol yes, alkanes no

The second consequence of the hydroxyl group is where castor oil will and will not dissolve, and this one has bench experience behind it rather than a measurement.

In our formulators' experience, castor blends with ethanol or isopropanol can stay clear at working ratios where an ordinary triglyceride would not, while blends with hydrocarbon emollients or silicones can go hazy or separate. We looked for a published miscibility measurement covering the pairs a cosmetic formulator actually uses and could not open one, so treat that as a hypothesis to test rather than a rule.

What can be quoted is the quantity underneath the behaviour: a hydroxyl value of 158.04 mg KOH/g against 0.66 to 4.22 for the four comparison oils in the same study. That supports the idea that castor can take part in far more hydrogen bonding than its neighbours. It does not measure ethanol miscibility, and it does not measure hydrocarbon compatibility.

The practical upshot is a habit rather than a rule: test the actual pair, at the actual ratio, water content and temperature, then leave it standing for a week at the temperatures the product will see. These failures appear on standing, not on mixing, which is why a five-minute compatibility check flatters a hydroalcoholic system and a seven-day one does not.

It is also why castor turns up more often in anhydrous and alcoholic systems than in a conventional cream oil phase — an argument developed in why your balm feels greasy when your cream does not.

At the bench: date the jar when you set a compatibility test standing, and look at it in a week under the same light. "It was clear when I mixed it" is not a result.

Soap: a small deliberate fraction

Soapmakers treat castor as a minority component of a blend rather than a base oil, and there is measurement behind the practice.

Borhan, Abd Gani and Shamsuddin in The Scientific World Journal (2014) built a D-optimal mixture design — a statistical way of laying out blends so that each component's effect can be separated — around six components, with castor oil constrained to 15 to 20 per cent by weight alongside virgin coconut oil, olive oil, palm oil, cocoa butter and okara. Across 19 formulations they measured bar hardness as penetration force. Hardness ranged from 239.8 to 593.1 N; their regression identified virgin coconut oil and castor oil as the two components with the greatest influence on it; and their optimum contained 16.241 per cent castor oil.

Note what was measured: hardness. Lather and conditioning appear in their discussion as citations to earlier literature, not as their own data — so the familiar claim that castor gives soap a creamy lather is not what this study demonstrates.

Transparent soap is where the record is thinnest of all. Patel and colleagues note that castor oil has been used to produce soaps and cite a 1981 transparent-soap patent; that is a mention of a use, not a measurement of clarity, lather or hardness, and we found no study measuring any of those against castor content. In our formulators' experience a single-digit to mid-teens percentage of the oil charge is a sensible starting range to test — but clarity, hardness and sweating have to be assessed in the complete formulation.

At the bench: if you are copying a soap recipe, note that the one measured optimum we can point to sat at 16 per cent of a six-oil blend, and that it was optimised for hardness alone. Everything else you have read about castor in soap is tradition until you test it.

Two lookalike names that are not the same material

This catches people because the two INCI names — INCI being the standard ingredient nomenclature that goes on the label — sit next to each other alphabetically:

• Hydrogenated Castor Oil is a hard, waxy structurant — a solid, used like a wax.
• PEG-40 Hydrogenated Castor Oil is an ethoxylated solubiliser, used to get fragrance into clear aqueous and hydroalcoholic systems.

They behave nothing alike, and the CIR assessment cited above covers castor oil and hydrogenated castor oil but not PEG-40 hydrogenated castor oil.

At the bench: read the whole INCI name on the invoice, not the first two words. Ordering the wrong one is a batch, not a typo.

The British part: where a balm stops being a cosmetic

This is the section that catches small UK brands, and it is not about chemistry.

Castor oil carries more folklore than almost any other oil on the bench — lash growth, brow growth, "castor oil packs", drawing things out, sorting things out. Some of that folklore, written on a label or an Instagram post, stops being folklore and becomes a medicinal claim, and at that point the product is no longer a cosmetic and the regulator is no longer the one you were expecting.

The MHRA's guidance on borderline products, updated 2 July 2026, gives the two limbs of the definition. A medicinal product is any substance or combination of substances "presented as having properties of preventing or treating disease in human beings", or any that "may be used by or administered to human beings with a view to restoring, correcting or modifying a physiological function by exerting a pharmacological, immunological or metabolic action, or making a medical diagnosis".

Read the first limb again. Presentation is enough. It is what you say about the product that can move it across the line, and the MHRA is explicit about looking at claims, explicit and implicit, across labelling, packaging, promotional literature, advertisements, websites, social media and customer reviews. The reviews on your own shop page are in scope.

Growth claims are the specific trap with this oil, and the evidence does not support them anyway. The TFOS Lifestyle workshop report on cosmetics and the ocular surface (2023) records castor oil as an eyelash conditioner present in around-eye cream, eyeliner, eyeshadow, glitter, makeup primer, makeup remover, mascara and serum — and states that "there is no research indicating that eyelash cilia or follicles are either positively or negatively impacted by topical application to the eyelash line". Anecdote on one side, no measurement on the other.

What remains sayable on a UK label — conditioning, appearance, the ordinary cosmetic vocabulary, and what each of those still needs behind it — is worked through in what a UK label can say about an oil.

One more caution belongs here, about the eye area generally. Couteau, Girard and Coiffard analysed 275 do-it-yourself recipes for eye and periocular products collected from French- and English-language blogs and social media (International Journal of Cosmetic Science, 2022) and concluded that quantities are very imprecise, that the substances proposed are not suited to the intended use, and that preservation of the finished products is not sufficiently guaranteed. It is a study of internet recipes rather than commercial products — which is exactly why it is worth reading before you copy one.

At the bench: write the label copy at the same time as the formula, not after it, and strike out every verb that describes an effect on hair, follicles or skin function. If the sentence would be a shame to lose, that is usually a sign it was doing the work a medicine claim does.

Heat, and a supply chain with one address

Two last practical matters.

Do not cook it. Patel and colleagues report that physical refining is not recommended for castor oil because of its sensitivity to heat, as it normally starts disintegrating above 150 °C, which can result in hydrolysis of the hydroxyl groups, and that pharmaceutical-grade castor oil is deodorised at a comparatively low 150–170 °C under high vacuum to avoid exactly that. Those are refinery conditions, and they do not establish a hot-process limit for your bench or prove preferential loss of hydroxyl groups in a balm pot. What they do tell you is that the group you are paying for is the heat-sensitive part of the molecule: set your processing limits from the supplier's guidance and your own stability testing, and melt gently.

And it comes from one place. Patel and colleagues note that global castor oil production is concentrated primarily in a small geographic region of Gujarat, in western India, with growers in the United States and South America facing challenges of production efficiency and oil quality. For a British brand that means a single-origin dependency on an ingredient with no functional substitute — nothing else has the hydroxyl group, so there is no drop-in replacement to switch to if a lot is late or a harvest is poor.

At the bench: qualify a second supplier before you need one, keep a retained sample of every delivery, and if a formula is going into production, hold enough stock to cover a reorder cycle. This is the one oil where "we will just use something else" is not an option.

What to check at the bench

When a formula containing castor misbehaves, ask whether the hydroxyl group is the cause or the alibi. It usually explains the tack and the solubility. It rarely explains the hardness.

Sources

• Patel V. R., Dumancas G. G., Kasi Viswanath L. C., Maples R., Subong B. J. Castor oil: properties, uses, and optimization of processing parameters in commercial production. Lipid Insights 9, 1–12, 2016. PubMed 27656091
• Cosmetic Ingredient Review Expert Panel. Final report on the safety assessment of Ricinus Communis (Castor) Seed Oil, Hydrogenated Castor Oil, Ricinoleic Acid and related ingredients. International Journal of Toxicology 26(Suppl 3), 31–77, 2007. PubMed 18080873
• Ciastowicz Ż., Pamuła R., Bobak Ł., Białowiec A. Characterization of vegetable oils for direct use in polyurethane-based adhesives: physicochemical and compatibility assessment. Materials 18(5), 918, 2025. PubMed 40077144
• Yalçın H., Toker O. S., Doğan M. Effect of oil type and fatty acid composition on dynamic and steady shear rheology of vegetable oils. Journal of Oleo Science 61(4), 181–187, 2012. PubMed 22450119
• Zhang M., Weiss R. G. Insights into the gelating abilities of ricinelaidic acid and its ammonium salts. ChemPhysChem 17(24), 4059–4067, 2016. PubMed 27862785
• Maktabi B., Liberatore M. W., Baki G. Meadowfoam seed oil as a natural dispersing agent for colorants in lipstick. International Journal of Cosmetic Science 43(4), 484–493, 2021. PubMed 34240435
• Borhan F. P., Abd Gani S. S., Shamsuddin R. The use of D-optimal mixture design in optimising okara soap formulation for stratum corneum application. The Scientific World Journal 2014, 173979, 2014. PubMed 25548777
• Budai L., Antal I., Klebovich I., Budai M. Natural oils and waxes: studies on stick bases. Journal of Cosmetic Science 63(2), 93–101, 2012. PubMed 22591561
• Ogunniyi D. S. Castor oil: a vital industrial raw material. Bioresource Technology 97(9), 1086–1091, 2006. PubMed 15919203
• Couteau C., Girard E., Coiffard L. Analysis of 275 DIY recipes for eye cosmetics and their possible safety issues. International Journal of Cosmetic Science 44(4), 403–413, 2022. PubMed 35396729
• Sullivan D. A. et al. TFOS Lifestyle: impact of cosmetics on the ocular surface. The Ocular Surface 29, 77–130, 2023. PubMed 37061220
• Medicines and Healthcare products Regulatory Agency. Borderline products: how to tell if your product is a medicine. gov.uk — updated 2 July 2026, consulted 22 September 2026.
• FAO/WHO Codex Alimentarius Commission. Standard for Named Vegetable Oils, CXS 210-1999 (2015 amendment). CXS_210e_2015 (PDF) — consulted 22 September 2026.