How Aroma Materials Are Extracted

Five methods matter. Steam and water distillation for robust plant material, mechanical expression for citrus peel, volatile solvent extraction for heat-sensitive flowers, supercritical or liquid CO2 for selective extraction without heat, and enfleurage, which is effectively obsolete. Each produces a chemically different product from the same plant.

What do essential oil, concrete, absolute, resinoid and tincture actually mean?

These are defined terms, not marketing words. ISO 9235, the international vocabulary standard for aromatic natural raw materials, defines each by the process that produces it. The definitions are worth reading closely, because almost every popular explanation of them is wrong in some detail.

Two consequences follow immediately. An “absolute” sold as a direct extraction of a flower is mislabelled, because by definition an absolute is washed out of a concrete, pomade, resinoid or supercritical extract with ethanol. And a tincture is the weakest form of natural material in the palette, not the strongest, because it is a dilute solution rather than a concentrate.

How does distillation work, and what does it do to the smell?

Steam is passed through, or water is boiled with, plant material in a still. Volatile compounds leave with the steam, the mixture is condensed, and because most aroma compounds are only slightly soluble in water, the oil separates and is drawn off. The aqueous phase left behind is the hydrosol or floral water.

A published account of Turkish rose production gives the working parameters: copper stills charged with 500 kg of flowers and 1,500 liters of warm water, distilled for about an hour and a half, with the condenser running at 35 to 45 °C. It produces a first oil and a second oil, blended afterward for the final product. The second oil comes from cohobation — redistilling the aqueous distillate to recover material that stayed in the water on the first pass.

That detail about water solubility is the most important thing distillation does to a smell, and rose demonstrates it better than anything else. Analytical work comparing the same Turkish roses by three methods found:

Phenylethyl alcohol is the honeyed, slightly sweet, unmistakably fresh-petal smell of a rose. It is also water-soluble, so it stays in the distillation water and drops from 43 percent of the living flower’s emission to 1.3 percent of the distilled oil. Geraniol and citronellol, which are more geranium-like and sharper, concentrate in its place. This is why rose otto smells lemony, waxy, faintly metallic and green-spicy, while rose absolute smells like a rose in a warm room. The plant is identical. The method rewrote the composition.

Dry distillation, the third route ISO recognizes, heats the material with no water at all. It produces birch tar and cade — smoky, tarry, leather-like materials that are pyrolysis products, not extracts. Nothing in a living birch smells like birch tar.

Distilled oils can also be worked further. Fractional distillation and rectification separate an oil into cuts by boiling point, which is how terpeneless citrus oils, individual isolates and deodorized fractions are made. Molecular distillation, run under high vacuum at low temperature, is used for heat-sensitive material such as orris butter.

Why are citrus oils pressed instead of distilled?

Because citrus oil is already sitting in reservoirs in the outer peel, and because heat ruins it. The aldehydes that give lemon oil its bright, cutting freshness degrade in a still; the resulting distilled oil is flatter and more terpenic.

Industrial expression uses several machine types, each with a different reputation. A pelatrice passes whole fruit between counter-rotating stainless steel drums under water sprays, forming an emulsion that is centrifuged. A sfumatrice carries already-juiced peel between chains and a fixed ribbed plate. An FMC in-line extractor, typically with four heads, takes oil and juice simultaneously as cups compress the fruit and cutters rupture the oil cells. A torchio forces rinds between two counter-rotating screws under regulated pressure. Trade reporting holds that sfumatrice oil is normally superior and commands a premium over FMC or pelatrice oil.

Expression has a safety consequence. Because it is a mechanical squeeze rather than a distillation, non-volatile peel constituents come along, including furocoumarins. Bergapten in expressed bergamot is phototoxic. The IFRA Standard for citrus oils and other furocoumarin-containing essential oils, in the Amendment 49 version currently published in IFRA’s Standards Library, states that “the total level of Bergapten in the consumer products should not exceed 0.0015% (15 ppm)” where the product is applied to skin exposed to UV light. That figure is being revised: IFRA has published the documentation behind a second leave-on threshold of 5 ppm, and the 52nd Amendment consolidates several natural-substance standards into one furocoumarin standard, so check the current Standards Library before relying on it. Bergapten-free bergamot exists as a separate commercial grade. Distilled citrus oils are not phototoxic in the same way, because the furocoumarins do not distill over. A bright citrus opening like Acqua di Parma Colonia is expressed-oil territory, and the reason its top is so short-lived is the same reason it is so vivid.

How does solvent extraction work, and why does it give a different smell?

Fresh plant material is washed with a volatile hydrocarbon solvent, historically petroleum ether or hexane, at or near ambient temperature. The solvent dissolves the aroma compounds along with plant waxes, pigments and other lipophilic material. The solvent is then stripped under vacuum, leaving the concrete: a waxy, colored, often semi-solid paste.

The concrete is then stirred with ethanol, which dissolves the aromatic fraction but not most of the waxes. The mixture is chilled, the precipitated waxes are filtered off, and the ethanol is removed. What remains is the absolute.

Two properties explain why absolutes smell closer to the living flower than distilled oils do. No heat, so thermally fragile compounds survive. And no water, so water-soluble compounds like phenylethyl alcohol are retained rather than left behind in a hydrosol. The tradeoff is that absolutes carry odorless ballast — in Turkish rose absolute, hexadecane at 28 to 36 percent — and residual solvent at trace levels, which is one reason absolutes are not permitted in some certification schemes.

This is the method behind almost every floral material a perfumer actually uses: jasmine, tuberose, orange flower, mimosa, narcissus, violet leaf, and rose absolute alongside rose otto. Floral-absolute construction is audible in something like Chanel No 5, where the rose-jasmine core is doing work that no distilled oil could do at that dose.

Is enfleurage still used?

Essentially no, and this is the single most repeated piece of misinformation about extraction.

Enfleurage works by diffusion into fat. In cold enfleurage, flowers are laid by hand onto a fat-coated glass plate held in a wooden frame called a chassis, left about 24 hours, then replaced with fresh flowers until the fat is saturated. The perfumed fat is a pomade, washed with ethanol to give a pomade absolute. In hot enfleurage, or maceration, flowers are digested in melted fat: a Grasse house describes the fat as 75 percent pork and 25 percent beef, held in a bain-marie at 40 to 60 °C for at least 24 hours with the flowers replaced daily.

That same Grasse house states that the technique was made obsolete around 1930, when volatile solvent extraction became reliable, and that it is “little used today”; it names recent enfleurage work on tuberose in Grasse and a Guerlain project using mango butter as revivals rather than continuing practice. That enfleurage has been superseded because it is inefficient and costly is uncontroversial, but the date of around 1930 rests on that single company page, so read it as the house’s own account rather than an established chronology. If a product description says its jasmine was obtained by enfleurage, that is a claim worth asking about: nothing in the public record suggests the material is available at scale.

What does CO₂ extraction do differently?

Carbon dioxide has a critical point at about 31 °C and 74 bar. Above that, it behaves as a supercritical fluid: gas-like in its ability to penetrate plant material, liquid-like in its ability to dissolve. Crucially, its solvent power can be tuned continuously by changing pressure and temperature, so the same plant can be extracted selectively.

The practical demonstration: volatile oils can be taken at relatively low pressures around 100 bar, while lipids require higher pressures, and more polar compounds require adding a modifier such as ethanol. Depressurize and the CO₂ leaves as a gas, so there is no solvent residue at all, and the whole operation runs near room temperature.

Whether it produces a better material depends on the plant, and the published jasmine comparison is instructive rather than flattering. A study on Jasminum grandiflorum found liquid CO₂ extraction gave a relatively fat-free product at 0.26 percent yield, against 0.35 percent for an n-pentane concrete, 0.27 percent for the absolute obtained from it, and 0.05 percent for steam distillation. CO₂ matched the absolute on yield while avoiding excess fats, recovering benzyl acetate, alpha-farnesene and indole well. It did not beat solvent extraction by a wide margin: it produced a cleaner, differently weighted material at a similar yield.

Where CO₂ genuinely wins is with materials that distillation destroys and solvent extraction muddies: vanilla, ginger, black pepper, hops, spices and resins. A CO₂ pepper extract smells recognizably like cracked pepper. Distilled pepper oil smells like pine and citrus with the pepper missing.

What are the real yields?

Yield figures in perfumery writing are the most corrupted data in the field. Here are the ones that trace to peer-reviewed papers or technical monographs. Everything in the table is a measured or published figure; the kilograms-per-kilogram columns are my arithmetic from those percentages, and are labelled as such.

Read the rose rows together and you have the answer to the most-repeated yield claim in perfumery. A technical monograph on oil-bearing rose cultivation states that “1 kg rose oil can be obtained from 3000-4000 kg of rose petals.” The measured yields imply roughly 2,040 to 2,860 kg, so the monograph figure is 40 to 50 percent higher than the measured one — that arithmetic is mine — probably because it describes petals as charged rather than whole flowers as harvested, and because real distilleries do not hit laboratory yields. That is an unusually good result for a widely repeated number. Most are worse.

How are aroma materials actually extracted from plants, method by method?

1. Decide whether the material survives heat. If it does, distillation is cheapest. If it does not — jasmine, tuberose, mimosa, narcissus — distillation is off the table before anything else is considered.
2. Decide whether the odor lives in an oil gland you can rupture. Citrus peel does, so it is pressed. Almost nothing else does.
3. Harvest into the window. Rose and jasmine are picked at dawn and processed the same day; delay costs yield and shifts composition. Orris is the opposite: it must be aged 36 months before the irones exist at all.
4. Prepare the charge. Distillation takes whole or comminuted material into the still. Solvent extraction takes fresh flowers into an extractor. Orris rhizomes are sun-dried about nine days first, losing roughly 70 percent of their fresh weight.
5. Run the separation. Steam through the still for about 90 minutes for rose; solvent wash at ambient temperature for flowers; pressure and temperature setpoints for CO₂.
6. Recover the first product. Oil separates from the hydrosol; solvent is stripped under vacuum to leave a concrete; CO₂ is depressurized to gas and vented.
7. Recover the second product where economics justify it. Cohobate the distillation water for a second oil; wash the concrete with ethanol, chill and filter to get the absolute.
8. Standardize. Blend lots against a reference, and rectify or fractionate where a cleaner or more consistent cut is needed.

Which extraction claims are wrong?

“Enfleurage is still used for jasmine and tuberose”

What is claimed: Fine perfumery still obtains its most delicate florals by enfleurage.

What the evidence shows: Standard reference accounts describe enfleurage as superseded by solvent extraction and supercritical fluid extraction on grounds of inefficiency and cost. A Grasse fragrance house puts the change at around 1930, when volatile solvent extraction became reliable, and says the technique is “little used today,” describing recent exceptions as new initiatives rather than continuing practice. The 1930 date is that one company’s account.

What people wrongly conclude: That enfleurage-derived materials are available at scale. They are not, and a product claiming them is making a claim about an artisanal revival, not about standard industry practice.

“It takes a thousand tonnes of iris to make two kilos of orris”

What is claimed: Vendor descriptions of orris cite ratios in the range of a thousand tonnes of rhizome per couple of kilograms of finished material.

What the evidence shows: The peer-reviewed field trial measured orris concrete yield at 0.045 to 0.055 percent of rhizome. Arithmetic, mine: at 0.05 percent, a thousand tonnes of rhizome gives about 500 kg of concrete. Even allowing that orris butter is a further distillation of the concrete and so a fraction of it, the vendor ratio is off by something like two orders of magnitude.

What people wrongly conclude: That orris is rare in a way that no arithmetic can express. It is genuinely one of the most expensive materials in perfumery — the driver is six years of capital tied up in a crop, plus about 6 kg of concrete per hectare per year — but the extreme ratios circulating in retail copy are not measurements.

“An absolute is a purer, more concentrated essential oil”

What is claimed: Absolutes are essential oils refined a step further.

What the evidence shows: ISO 9235 defines an essential oil by distillation, expression or dry distillation only; an absolute is an ethanol extract of a concrete, pomade, resinoid or supercritical extract. They are different product classes from different processes. And absolutes are not purer: Turkish rose absolute was measured at 28 to 36 percent hexadecane, an odorless alkane.

What people wrongly conclude: That absolute is a quality grade. It is a process designation, and for some plants — rose, in particular — the absolute and the distilled oil are usefully different materials that a perfumer uses for different jobs.

“Essential oil captures the smell of the flower”

What is claimed: Distillation gives you the plant’s true scent.

What the evidence shows: Living rose headspace is 43.2 percent phenylethyl alcohol. Distilled rose oil is 1.3 percent. The compound most responsible for the fresh-petal smell is largely absent from the distilled oil because it is water-soluble.

What people wrongly conclude: That essential oil is the reference against which other extracts are judged. For most flowers it is the least faithful of the available options.

“CO₂ extraction is always superior”

What is claimed: Supercritical extraction is a strictly better technology.

What the evidence shows: On jasmine it matched the absolute on yield (0.26 versus 0.27 percent) while avoiding excess fats. Better for some purposes, not all: it produces a differently weighted material, and for many plants the solvent-extracted absolute remains the reference perfumery material.

What people wrongly conclude: That a CO₂ label signals quality. It signals a process choice suited to particular materials, most convincingly spices, resins, vanilla and hops.

The honest limits of this article

Yields are the weak point of every article on this subject, including this one. Published percentages come from specific cultivars, seasons, equipment and laboratories; commercial extractors achieve lower numbers than research stills, and do not publish them. Any kilograms-per-kilogram figure derived from a percentage carries that uncertainty compounded by whatever the original study meant by “flowers” — whole flowers, petals only, fresh or partly wilted — which is frequently not stated. Flower-count claims of the “it takes eight million blossoms” type cannot be verified without a stated average weight per flower, which is almost never given, so this article does not repeat any of them. Where a figure here is arithmetic rather than measurement, it says so.

What this changes about what you smell

The immediate payoff is that “rose” on a box tells you almost nothing, and you now know what question replaces it. Rose otto and rose absolute are different materials from the same plant: otto reads lemony, waxy, sharp and slightly metallic; absolute reads honeyed, jammy, deeper, closer to a cut flower in a warm room. A fragrance built on one does not smell like a fragrance built on the other, and the note list will not distinguish them.

The same logic applies down the list. A citrus top that vanishes in fifteen minutes is expressed oil behaving normally, not a fault. A pepper note that smells piney is distilled; one that smells like a pepper grinder is a CO₂ extract. A vanilla that smells woody and slightly smoky rather than sweet is closer to a real bean extract than a vanillin-led accord. And an oud note in a widely distributed fragrance such as Tom Ford Oud Wood is largely reconstruction work, because the extraction economics of real agarwood do not scale.

What our own catalogue shows

Two figures from our listings show how little of this reaches a label. About 334 mention rose; of those, eight specify damascena or damask, six Turkish rose, five May rose, two rose absolute, none rose otto. About 112 mention oud or agarwood, at a median of roughly $90 per 100 ml against $68 catalogue-wide — a premium buying a reconstruction, not distilled oud oil, which at commercial volumes does not exist. Both counts reflect which copy names a material, not which formulas use it.

Bottles to try this on

For iris, the method determines almost everything: Prada Infusion d’Iris leans on the cool, dry, faintly floury side of the material, and it is worth smelling next to a warm, powdery-vanillic construction like Guerlain Shalimar to hear how differently powder can be built. For a pure floral-absolute reference against a synthetic-led one, compare a jasmine-heavy classical structure with something like Dior Sauvage Eau de Parfum, which contains essentially none of the processes described in this article.

Related reading

• how perfume is made, from field to bottle — six years in the field, a few hours on the line
• natural versus synthetic materials, honestly compared — behavior, supply and the quality myth
• how a formula gets written — briefs, benchmarks, submissions and formula cards
• Lattafa, Armaf, Rasasi and the Gulf houses — the houses, the price ladder and the inspired-by trade

Common questions

What is the difference between an essential oil and an absolute?

Process. ISO 9235 defines an essential oil as a product of steam distillation, mechanical expression of citrus peel, or dry distillation. An absolute is obtained by washing a concrete, pomade, resinoid or supercritical extract with ethanol. They are different product classes, and for flowers like jasmine only the absolute route works at all.

What is a concrete in perfumery?

A concrete is the first product of solvent extraction: fresh plant material washed with a hydrocarbon solvent, which is then stripped off, leaving a waxy colored paste containing aroma compounds plus plant waxes and pigments. Jasmine concrete yields are about 0.26 to 0.30 percent of flower weight. Washing it with ethanol gives the absolute.

Why does rose oil smell different from rose absolute?

Because distillation loses the water-soluble compounds. Analytical work on the same Turkish roses measured phenylethyl alcohol at 43.2 percent of the living flower’s emission, 1.3 percent of the distilled oil and 35 to 38 percent of the absolute. That compound carries the honeyed fresh-petal character, so the oil reads sharper and more geranium-like.

Is enfleurage still used to make perfume?

Not commercially. Standard reference accounts describe it as superseded by solvent extraction and supercritical extraction because it is inefficient and costly. A Grasse fragrance house puts the change at around 1930 and says the technique is little used today, naming recent tuberose work in Grasse as a revival. That 1930 date is one company’s account rather than an established chronology.

Why are citrus oils cold pressed instead of distilled?

Because the oil already sits in glands in the outer peel and can be squeezed out mechanically, and because heat degrades the aldehydes responsible for citrus brightness. Distilled lemon oil is flatter and more terpenic. The tradeoff is that expression also carries over phototoxic furocoumarins such as bergapten in bergamot, which IFRA restricts.

What does CO2 extraction actually do?

Carbon dioxide above about 31 degrees C and 74 bar behaves as a supercritical fluid that penetrates plant material like a gas and dissolves like a liquid, with solvent power tunable by pressure and temperature. It leaves no residue, since the CO2 is simply vented as gas, and it runs near room temperature so heat-sensitive compounds survive.

How many roses does it take to make rose oil?

Published yields for Rosa damascena are 0.035 to 0.049 percent of fresh flower weight, which works out to roughly 2,040 to 2,860 kg of flowers per kilogram of oil by my arithmetic. The frequently repeated figure of 3,000 to 4,000 kg comes from a technical monograph describing petals rather than whole flowers, and is roughly 40 to 50 percent higher than the measured range.

What is a tincture in perfumery?

ISO 9235 defines it as a solution obtained by macerating a natural raw material in ethanol or water. The raw material simply sits in the solvent and the solution is the product. Nothing is concentrated, so a tincture is a weak material used for specific effects, not a concentrated extract.

Which extraction method gives the most authentic smell?

It depends on the plant. For flowers, solvent extraction and CO2 extraction come closest to the living smell because neither uses heat or water. For citrus peel, expression is the most faithful. For spices and resins, CO2 extraction. For robust herbs and woods, distillation is both faithful and cheapest.

What is cohobation in rose distillation?

Redistilling the aqueous distillate to recover aroma compounds that dissolved in the water on the first pass. A Turkish production account describes it producing a second oil, which is then blended with the first oil for the final product. It partly compensates for the loss of water-soluble compounds like phenylethyl alcohol.