Carrier oils compared: how to choose the liquid oils in a balm

How to choose carrier oils for balms on four axes: oxidative stability, sensory weight, cost and functional extras, with three worked blend formulas.

Most balm recipes name an oil and move on. That is the wrong level of detail, because the liquid oil fraction is usually 45 to 70% of a balm's weight and it decides almost everything you notice: how long the product lasts before it smells of crayons, whether it feels like a film or a slick, what it costs, and whether a tinted version has any shine.

This page is the reasoning companion to the numbers. The full carrier oil comparison table holds the dataset. Here we deal with how to choose.

The four axes that actually decide an oil

Every other property people list (colour, provenance, marketing story, vitamin content) is downstream of these four, or too small at balm use levels to matter.

1. Oxidative stability. How fast the oil goes rancid, which sets the shelf life of the finished balm and is the single most common reason a home batch fails.
2. Sensory weight and absorption. Whether the film reads as dry, light, cushioned or greasy, and how long it stays detectable.
3. Cost and availability. Price per kilogram, minimum order size, and whether you can buy the same specification again in six months.
4. Functional extras. Gloss, cling, film formation, pigment wetting, natural odour, natural colour. These are the reasons to accept a bad score on one of the first three.

Score a candidate out of five on each axis before buying. An oil mediocre on all four has no reason to be in your formula.

Axis one: oxidative stability and the oleic/linoleic trade-off

Vegetable oils are mostly triglycerides: three fatty acids esterified to a glycerol backbone. Oxidation starts by abstracting a hydrogen, and it takes the weakest one available, which is a bis-allylic hydrogen on a carbon sitting between two double bonds. Linoleic acid (C18:2) has one such position, alpha-linolenic (C18:3) has two, and oleic acid (C18:1), despite being unsaturated, has none at all. That is why oleic behaves almost like a saturated fat on a balm's timescale, and why the useful number is not total unsaturation but linoleic percent plus twice linolenic percent.

How much faster the polyunsaturates go depends on what you measure. Ratios for oleic to linoleic to linolenic run 1:40-50:100 by oxygen uptake, 1:12:25 by hydroperoxide formation and 1:70:265 by propagation rate constants. Quote the range, not one number, and ignore the 1:100:1200:2500 figure circulating in craft sources, which has no traceable primary source. The consequence stands either way: high-oleic sunflower (index under 18) far outlasts standard sunflower (index 48 to 75), and rosehip, with linolenic reported anywhere from 16.6 to 51%, is the least stable oil in common use. The derivation is in rancidity and oxidation.

Linoleic acid is not merely a liability. It is the fatty acid of ceramide 1 linoleate in the intercellular lipid lamellae, and skin with a disordered barrier tends to be linoleate-poor, which has made high-linoleic oils popular for barrier claims. The observation is about skin lipid composition, not a demonstration that painting linoleic oil on the surface rebuilds ceramides. High-linoleic oils are pleasant, light, plausibly useful and short-lived, so they belong at 10 to 30% of an oil phase rather than as the base.

Three further points that people get wrong. First, a nominally stable oil that arrives already half-oxidised is not stable: life is counted from pressing, not from your purchase date, and a certificate showing a peroxide value of 8 meq/kg is legal but not fresh, since genuinely fresh oil tests at 0.1 to 1. Second, antioxidants shift the curve, they do not redraw it, and the dose is much smaller than usually quoted: mixed tocopherols at 0.02 to 0.1% of the oil phase, because alpha-tocopherol turns pro-oxidant above about 100 ppm. That correction is set out in vitamin E and antioxidants. Third, heat during making costs you life, so hold the melt at 70 to 80 C and never above 85 C for long.

What is actually known about shelf life

The shelf lives circulating in every carrier oil table, this site's included, are supplier estimates with no stated method, temperature, packaging or endpoint. Useful as guidance, worthless as evidence. Two things have been measured properly.

Jojoba lasted about nine times longer than rosehip, and the ranking follows polyunsaturate content almost exactly. Jojoba has also been measured at 68.6 to 71.2 hours at 110 C and 51 to 229 days at 60 C, against 1 to 29 days for every other oil tested.

The second finding matters more to a maker. In a 270-day real-time study of hemp seed oil (Bonazza et al., Molecules, 2024), the same oil in polypropylene at 25 C in diffused light reached 40 meq/kg of peroxide by day 150, nearly three times the Codex limit for a cold pressed oil, while in amber glass with foil at 10 C in the dark it stayed under 10 meq/kg for the full 270 days. Packaging and temperature beat oil choice. Before paying four times as much for a stabler oil, move the stock bottle out of the light and into the cold, and see the packaging guide.

Axis two: sensory weight, and why "dry oils" feel dry

"Dry" and "greasy" are not chemical categories, they are descriptions of how quickly a film stops being perceptible. Three properties drive it.

Molecular size. Smaller molecules spread further and thin faster. Caprylic/capric triglyceride is built from C8 and C10 acids rather than C18, so it spreads to a thinner film for the same weight applied. Squalane, a C30 branched hydrocarbon with no ester groups, does the same by being non-polar and low in viscosity.

Polarity. Polar oils such as castor, with its hydroxyl group, interact strongly with the skin surface and with each other. They cling, resist wiping and read as substantial. Non-polar oils sit lightly and are removed easily.

Viscosity. Directly perceived as body. Castor at roughly 950 centipoise at room temperature and avocado near 60 feel completely different at identical percentages.

Absorption is mostly an illusion. Very little oil crosses the stratum corneum. What you read as absorption is the film spreading, thinning and wicking into the surface topography until it drops below the sensory threshold, which is why "fast absorbing" and "light" describe the same physics. The distinction between sitting on top and softening the surface is set out in occlusive, emollient and humectant.

Axis three: cost, supply and the boring constraints

Cost per kilogram spans two orders of magnitude between high-oleic sunflower and pomegranate seed oil. The arithmetic is unforgiving: at 60% of the formula, an oil ten times the price of your base multiplies the ingredient cost about sixfold. The defensible pattern is a cheap, stable, neutral base at 50 to 70% of the oil phase carrying a small percentage of something expensive that does a specific job.

Supply consistency matters more than makers expect. Profiles shift with cultivar, season, pressing and refining, so two batches of the same nominal oil can differ by ten percentage points on oleic content. If your formula depends on a narrow window, specify it, ask for a certificate of analysis, and keep it with the batch record.

Two constraints sit alongside cost. Allergen exposure: almond, hazelnut, macadamia, walnut and other tree nut oils are a labelling and liability issue in lip and child products, whatever their skin feel. And odour: sesame, hemp, unrefined olive, neem and black cumin carry their own smell into the balm at levels no flavour will mask.

Axis four: functional extras worth paying for

Two have their own pages: jojoba, a liquid wax rather than an oil, and castor oil, the only reliable route to genuine shine. The third material people reach for, coconut, is more complicated than either camp admits: see coconut oil in balms.

How blending actually works

Blending is close to linear for the properties you care about. Fatty acid profile is exactly linear on a weight basis, so the index of a blend is the weighted average of its parts: mix 50 g of an 80% oleic oil with 50 g of a 20% oleic oil and you have a 50% oleic blend. Viscosity is closer to linear in the logarithm, so a little of a very viscous oil moves the blend less than you expect. Oxidative stability is the awkward one: it is dominated by the least stable component, because the peroxides forming in it accelerate oxidation of everything else. Ten per cent of a rancid oil takes the whole jar with it.

Three working rules follow. Set the base first at 50 to 70% of the oil phase, on stability and cost. Add feel-modifiers at 15 to 35%, functional extras last at 5 to 15%, and keep the least stable component under 15% unless the product will be used within three months. For the weighted profile of your own blend, use the oil blend calculator.

Three worked blends

All three are oil phases only, weight percent of the oil phase, totalling 100. Drop them into a normal balm structure: for a twist tube, 10 to 20% wax and 15 to 30% butters with the rest oil phase, per lip balm ratios.

Add 0.05% mixed tocopherol on the oil phase and this is about as durable as a plant-oil balm gets. We do not put a month figure on it, because no method exists that would justify one.

Blend B is deliberately expensive. If cost matters, swap the squalane for more fractionated coconut and accept a slicker finish.

Blend C contains a tree nut oil, so it is unsuitable for a general-audience lip product without prominent labelling. Substitute apricot kernel for the almond: the feel is close and apricot is not a tree nut allergen in the regulatory sense.

Sanity checks before you commit a blend

Then make 50 g, fill three tubes and a tin, and put one tube somewhere warm. A blend correct on paper can still be draggy, slippery or slow to set. Accelerated storage at 40 to 45 C tells you quickly whether the balm holds together, though it will not give you a date: see shelf life testing.

What to do next

Starting from scratch, buy three oils, not eight: high-oleic sunflower, jojoba and castor make a competent lip balm, salve and body balm between them and will still be usable in a year. Add apricot kernel for lightness and avocado for weight. Everything beyond that should answer a specific complaint about a specific batch. If you are using the kitchen cupboard instead, read kitchen oils in balms first.

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