Why one oil feels light and another feels heavy on the skin

Why one oil feels light and another feels heavy on the skin

Two bottles on the bench. The supplier calls one of them light and dry-feel and the other one rich and cushioning. You put a drop of the "light" one on your cheek and half an hour later it is still sitting there with a shine on it, while the "rich" one a friend swears by disappeared in seconds on her.

Neither of you is wrong, and neither is the catalogue. The problem is that light, dry, rich, fast and cushioning are adjectives. They are not measurements, they are not comparable between suppliers, and the oil one formulator calls light is the one another finds greasy on the third pass.

There are two numbers underneath those adjectives that are measurements: spreading value and polarity. This article is about what each one actually measures, what it can settle for you at the bench, and the four or five places where reading it wrongly costs you a batch. How an oil behaves once it is inside a finished cream, with an emulsifier and a thickener acting on it, is a different question with a different answer — that one is why your cream feels greasy.

Spreading value: a number that means "this one felt dry"

Start with the feeling, because the number was built to match it.

Put a fixed drop of neat oil on a skin-like surface, leave it alone, and come back after ten minutes. A thin, fast oil will have crept out into a wide patch. A heavy one will still be more or less where you put it. Measure the area of the patch in square millimetres and you have the spreading value — reported as mm²/10 min, an area at a stated moment rather than a final state. That is why the unit has a clock in it.

The protocol everyone quotes comes from Mihaela Gorcea and Donna Laura at International Specialty Products, in Cosmetics & Toiletries (print issue 125(12), December 2010; posted online 31 May 2013). They used a synthetic membrane made to mimic the surface topography, pH, surface tension and wetting properties of human skin, cut to roughly 4 × 15 cm and hydrated for 24 hours over a 15% w/w glycerin solution; a 10 µL sample placed by electronic pipette; the area read after 10 minutes, in triplicate, six measurements in all.

You do not need to own any of that. You need the one consequence of it: a spreading value is only a spreading value under the method that produced it. Another house, another membrane, another volume, another ten minutes — and the figure moves.

At the bench: when you write a spreading value into your ingredient notes, write the source and the method on the same line. A catalogue figure with no method behind it is a marketing adjective with a number attached, and it cannot be compared with anything.

The published fast–medium–slow scale, and what the panel said

Gorcea and Laura tested four emollient esters chosen to span a range of molecular weight and branching, then put the same four neat materials on the volar forearms of a twelve-person panel. This is the clearest published fast/medium/slow scale with its method attached that we could open, so here it is — feel first, figure second.

Two things there matter more than the numbers themselves.

The first is that everything agreed. The rank order out of the membrane matched the rank order out of the contact angle, the surface tension and the twelve people. The heaviest ester had the highest initial contact angle (80 degrees), the highest surface tension (33 mN/m) and a viscosity of 83 cP; the middle one sat in the middle on every axis, with an initial contact angle of 69.4 degrees. When instruments and hands agree this cleanly, the instruments are measuring something real.

The second is the limit written into the study. Every one of those four is a synthetic ester. A plant oil is a triglyceride — chemically a triester, but not one of these. Reading an ester scale straight across onto sunflower or avocado is the first and most common way this number goes wrong.

At the bench: use this table as a shape, not as a lookup. It tells you what a four-step ladder from "vanishes" to "stays" looks like; it does not tell you where your rosehip oil sits on it.

What actually drives the number: viscosity, mostly

The largest study here is by Douguet and colleagues in Colloids and Surfaces B: Biointerfaces (2017), who characterised 53 emollients — esters, silicones, vegetable oils and mineral oils — for viscosity, surface tension, density and spreadability, the last by two independent routes on Vitro-skin, an artificial porous skin substitute.

Their contact-angle work explains what the ten minutes are actually made of. Spreading happens in two phases: first the drop spreads on its own, driven by wetting; then absorption and migration into the porous substrate takes over as the driver of how far the covered area extends. So a spreading value folds two different physical processes into one figure, and the porosity of whatever you put the drop on is part of the answer.

Statistically, the thing that dominates is viscosity. Douguet found viscosity had a major impact whatever the chemical type of the emollient, and was able to write spreading value as a linear function of viscosity — but separately for each of the four families, because esters, silicones, vegetable oils and mineral oils each had their own line. Even inside the ester family the authors split "high viscosity esters", where viscosity is the main factor, from "low viscosity esters", where structural details (mono- versus diester, saturated versus unsaturated chain, linear versus branched) have to be weighed alongside it.

Bom and colleagues (2021) reached a compatible answer from a different direction. Modelling the spreading of sustainable alternatives to petrolatum and dimethicone, they threw eight candidate predictors at the problem — spreading value, apparent viscosity, density, saponification value, iodine value, peroxide value, acid value and melting range — and their two optimised models kept viscosity and density, with melting range added in the petrolatum case. Saponification and iodine values, the two that most often appear on a plant-oil specification sheet, were tested and not retained.

At the bench: viscosity is a legitimate first sort within one chemical family — among your esters, or among your plant oils, not across both. And the numbers on a plant-oil spec sheet that look like they should predict feel (saponification value, iodine value) did not survive the modelling.

"Spreadability" is at least five different measurements

Al-Barghouthy and colleagues mapped this field in Gels (2025): a scoping review that screened 211 records, included 14 studies, then ran all five techniques — parallel-plate, slip-and-drag, flow-curve and amplitude-sweep rheometry (rheometry being the measurement of how a material flows and deforms), texture analysis and frictiometry — on ten commercial formulations.

Their opening statement is that there is no standardised method for assessing spreadability across pharmaceutical and cosmetic applications. Their conclusion is that no single method universally captures it. Texture analysis and amplitude-sweep rheometry correlated at r = 0.74; flow-curve yield stress correlated negatively with parallel-plate spreadability at r = −0.796; frictiometry varied significantly with formulation type. They recommend a tiered approach rather than one number.

At the bench: never put two spreading figures from two suppliers in the same comparison, and never compare a "spreadability" from a rheometer with a "spreading value" from a membrane. They are different measurements wearing the same word.

Polarity: the second axis, and what it is for

Spreading value tells you how far an oil goes. Polarity is about what it will dissolve and how it sits against water. It is the axis people skip, and it is the one that decides whether your UV filter stays in solution until the jar is empty.

Unlike spreading value, polarity has several rival measurements that do not fully agree. El-Mahrab-Robert and colleagues in the International Journal of Pharmaceutics (2008) compared three of them head to head on eight oils: the dielectric requirement for solubilisation, interfacial tension, and chromatographic analysis by HPLC. Their result is a ranking of methods rather than a table of oils — HPLC and interfacial tension were complementary to each other and superior to the solubilisation method for classifying oils by polarity.

Dielectric-type scales are the other family in common use and the one to read carefully. Gorcea and Laura report 5.35 for diisopropyl adipate, the highest of their four and therefore the most polar. Read as a relative permittivity that figure is a dimensionless ratio, so the unit as printed in the article is loose; the ranking is the usable part. That is the general rule for polarity scales: trust the order, distrust the absolute number, and never compare a dielectric figure from one source with an interfacial tension from another as though they were one quantity.

Where polarity earns its keep is solubility. A sunscreen's oil phase has to hold its organic UV filters in solution through manufacture and the whole of the product's shelf life, and crystallisation may change how the filter is distributed and how the product performs — which is something to assess in the finished formulation, not predict from a table. Gorcea and Laura name the type, use level and solubility of the UV filter in the ester among the parameters that decide emollient use level in the first place (they put typical skin-care emollient use at 3–20% w/w), and state the design rule plainly: matching the dielectric constants of emollient esters, and thus their polarity, is important to assure optimal solubilisation and to improve formulation stability.

Goussard, Aubry and Nardello-Rataj make the same point at review level in Advances in Colloid and Interface Science (2022): since no single carbon-based emollient imitates the non-sticky, dry feel of silicone oils, it is judicious to blend alkanes and esters, because the mixture improves both the sensory properties and the solubilising properties towards polar ingredients — they name sun filters, antioxidants and fragrances. The alkane brings the feel, the ester brings the solvency, neither does both.

At the bench: if your formula carries a UV filter, an oil-soluble antioxidant or a fragrance, keep at least one polar ester at the fast end of the blend. The same applies to the unsaponifiable fraction an unrefined plant oil brings with it — the minor components that are not fats at all, tocopherols included. Those have to stay in solution too.

Polarity is not required HLB, and this one is expensive

Required HLB (hydrophilic–lipophilic balance — the emulsifier balance a given oil phase needs) is an empirical property of a specific oil phase. It is not derivable from a structure, a polarity figure or a spreading value, and the spread between oils is wide.

Lee and Yoon in the Journal of Cosmetic Science (2020) determined required HLB for four materials by making O/W emulsions with lipophilic and hydrophilic surfactants at varied ratios, measuring droplet size and turbidity, and taking the HLB of the emulsion with minimum dispersion ratio, minimum droplet size and maximum turbidity. They report 14.75–14.90 for Citrus unshiu fruit oil, 15.35–15.40 for Citrus unshiu peel oil, 6.30–7.06 for horse fat and 5.94–6.30 for Camellia japonica seed oil.

Under one method, in one laboratory, the camellia seed oil and the citrus peel oil came out about nine HLB units apart. That is the size of the trap. A light, fast ester gets swapped in late to fix the feel, the feel is fixed — and the required HLB of the phase has moved underneath it. Six weeks later the emulsion starts creaming and nobody connects it to the sensory tweak.

At the bench: changing the oils and re-checking the emulsifier are one decision, not two. Make the swap, then re-check the phase.

Building a cascade — and what it can and cannot buy you

The cascade idea is simple: pick three or four emollients at different points on one method-consistent spreading scale, so that the film changes character as the fastest components leave the surface. A fast material to carry the rub-out and disappear, a medium one to bridge, a slow substantive one to be the after-feel. Polarity is chosen alongside speed, because the fast end usually has to carry whatever needs a solvent.

Direct experimental evidence on cascades was thin until recently. Gerken and colleagues in the International Journal of Cosmetic Science (2026) built 32 O/W emulsions around one fixed emulsifier system at two loads (2.5% and 5%), varying the emollient blend, its ratio and the total oil load (9% or 18%), and put them to 50 untrained panellists using a check-all-that-apply survey. Their multiple factor analysis put appearance at 47.73% and afterfeel at 24.66% of the perceptual structure, both driven by the emollient cascades — while pick-up, rub-out and immediate skin feel contributed marginally. The emulsifiers set the prevailing sensory profile; the cascade shaped static afterfeel and visual cues.

At the bench: if the brief is "less greasy at ten minutes" or "less shiny in the mirror", the cascade is your lever. If the brief is about the first three seconds — how it picks up, how it rubs out — changing oils is the slow way round, and the emulsifier will overrule you. Build the cascade from materials whose spreading values were measured by the same house under the same protocol, because that is the only comparison that holds.

Where these numbers mislead

Comparing across methods. Covered above, and it is the commonest error: supplier values are measured under supplier protocols.

Assuming viscosity or lipid content will stand in for feel in a finished product. Huber and colleagues in Skin Pharmacology and Physiology (2025) had nine trained panellists evaluate ten emollient products against 18 predefined sensory attributes, with viscosity on a rotational rheometer, and found lipid content, viscosity and galenic product format are not generally indicative of sensory product attributes. That is not in conflict with Douguet: one was measured on neat emollients, the other on finished products.

Reading one axis as the whole picture. Parente, Gámbaro and Solana (2005) characterised eight liquid emollients — mineral oil, sunflower oil, squalane, decyl oleate, isopropyl myristate, octyldodecanol, dimethicone and cyclomethicone — instrumentally and with a 14-assessor panel. Physicochemically the eight sorted into three groups, the silicones distinctly separate; sensorially into four, with isopropyl myristate also pulling away. Partial least squares showed sensory attributes could be well predicted from the instrumental measures — but it took several of them together.

Ignoring what the oil actually is. Nothing in a spreading value tells you fatty-acid composition, oxidative behaviour or minor components. Two oils with identical spreading values can differ in all three — those axes are the subject of plant oils by fatty-acid group and the pieces on oleic and linoleic oils. Spreading value is a physical descriptor with a sensory correlate; it is not a quality grade.

If you sell in Britain, the adjective becomes a claim

This is the part that catches small British brands, and it is not about chemistry.

The moment a supplier's word — "dry oil", "fast-absorbing", "lightweight" — moves from their catalogue onto your label, your listing or your Instagram post, it stops being trade shorthand and becomes a cosmetic claim about your product. Article 20 of the UK Cosmetics Regulation, as it stands in retained GB law, says text, names, trade marks, pictures and other signs "shall not be used to imply that these products have characteristics or functions which they do not have" — and Article 20(2), as substituted for GB, puts the duty on a named person: "A responsible person must ensure that the wording of any claim in relation to a cosmetic product is in compliance with the common criteria set out in the Annex to Commission Regulation (EU) No 655/2013". What those criteria demand of an ingredient-led sentence is set out in what a UK label can say about an oil.

Two more GB-specific consequences of changing an oil, both from the government's own guidance on making cosmetic products available in Great Britain:

• Your product's safety "must be checked by a qualified safety assessor" before you make it available. That assessment was written about a particular oil phase. Re-cutting your cascade is a formulation change, not a sensory tweak.
• The Responsible Person — who must have a UK established address, and a mail-forwarding service or PO box does not count — keeps the Product Information File, which has to include the product safety report and "evidence for the cosmetic product's effects". If the effect you are claiming is how the product feels, that evidence is what you are expected to hold.

And a practical British note: UK suppliers price emollients in pounds per 100 ml or per kilo, and a great many of them reprint the same handful of supplier tables without the method attached. The spreading figure on two different British websites is quite often the same figure copied twice — which does not make it comparable with a third one measured somewhere else.

What to do with the numbers at the bench

The habit that pays is narrow and dull: record the method beside every spreading value you keep, never mix scales in one comparison, and when you change one emollient in a cascade, re-measure the phase rather than reasoning from the catalogue. Physical data inform the choice. They promise nothing about the finished product, which has to be validated on each formulation and each production batch.

Sources

• Gorcea M., Laura D. Evaluating the physiochemical properties of emollient esters for cosmetic use. Cosmetics & Toiletries 125(12), December 2010; posted online 31 May 2013. Supplier-authored technical article (International Specialty Products); cited for its stated method and the values reported under it. cosmeticsandtoiletries.com
• Douguet M. et al. Spreading properties of cosmetic emollients: use of synthetic skin surface to elucidate structural effect. Colloids and Surfaces B: Biointerfaces 154, 307–314, 2017. PubMed 28371727
• Bom S. et al. A mathematical modeling strategy to predict the spreading behavior on skin of sustainable alternatives to personal care emollients. Colloids and Surfaces B: Biointerfaces 205, 111865, 2021. PubMed 34044331
• Al-Barghouthy E. Y. et al. Comparative evaluation of spreadability measurement methods for topical semisolid formulations: a scoping review. Gels 11(12), 1006, 2025. PubMed 41441162
• El-Mahrab-Robert M. et al. Assessment of oil polarity: comparison of evaluation methods. International Journal of Pharmaceutics 348(1–2), 89–94, 2008. PubMed 17728082
• Goussard V., Aubry J. M., Nardello-Rataj V. Bio-based alternatives to volatile silicones. Advances in Colloid and Interface Science 304, 102679, 2022. PubMed 35512559
• Lee Y. Y., Yoon K. S. Determination of required HLB values for Citrus unshiu fruit oil, Citrus unshiu peel oil, horse fat and Camellia japonica seed oil. Journal of Cosmetic Science 71(6), 411–424, 2020. PubMed 33413785
• Gerken A. et al. Sensory contribution of emollient cascades in skin creams evaluated by CATA analysis. International Journal of Cosmetic Science, 2026. PubMed 42554169
• Huber P. et al. How to choose an emollient? Pharmaceutical and sensory attributes for product selection. Skin Pharmacology and Physiology 38(1–2), 1–9, 2025. PubMed 39832490
• Parente M. E., Gámbaro A., Solana G. Study of sensory properties of emollients used in cosmetics and their correlation with physicochemical properties. Journal of Cosmetic Science 56(3), 175–182, 2005. PubMed 16116522
• Regulation (EC) No 1223/2009, Article 20 (Product claims), as retained GB law. legislation.gov.uk — consulted 22 September 2026.
• Office for Product Safety and Standards. Making cosmetic products available to consumers in Great Britain. gov.uk — consulted 22 September 2026.