How plant material becomes essential oil — the machine, the physics, and why the method decides what is in the bottle.
~11 min read
Every bottle on your shelf is one operation done at scale. Heat a plant in the company of water, catch what leaves, let it settle. The top layer is the oil.
Distillation is separation by volatility: raise the temperature, the volatile parts depart, cool that vapour elsewhere and the mixture is sorted without being touched. Spirits, petroleum and your oil shelf all come out of that idea.
The apparatus is an alembic: a body holding the material, a domed head gathering the rising vapour, a long spout running down so the vapour cools and arrives as liquid. Three parts, one of them cold. A thousand years on the vessel is steel rather than copper and holds tonnes rather than kilograms, but the principle has not moved.
The word reached English from the Arabic al-anbiq, which took it in turn from the Greek ambix — the name of the head itself, the part that does the collecting.
Craftsmen in Alexandria were distilling in the first centuries of the common era, but the instrument as later centuries used it was refined in the Arabic- and Persian-speaking world, and the product that drove it was rose water. Physicians there were making it at scale by the ninth century. The improvement usually credited to Ibn Sina — Avicenna, writing around the turn of the eleventh century — is the coil-cooled condenser and the use of steam in place of direct boiling. Historians argue over how much of that belongs to him rather than to the workshops he described. What is not in dispute is that the method was already industrial, and that the oil floating on top of the rose water was at first the by-product rather than the point.
Diagram to come
Two stills side by side at equal size, an eleventh-century copper alembic on the left and a modern stainless steel steam still on the right, drawn in the same line weight so neither reads as the more serious machine. The same four functional parts are numbered identically on both — the vessel holding the plant material, the head or dome where vapour collects, the descending cooled pipe, the receiver where the two liquids part — with faint connecting lines across the gap so part one can be traced to part one. The real differences must still be visible: the alembic is heated from beneath over a fire, the modern still takes steam from a separate boiler through an inlet under a perforated grate. The single point of the drawing is that the parts correspond.
A still has four functional parts: a vessel for the material, a way of getting steam into it, something cold for the vapour to meet, and somewhere for two liquids to part. What essential oils are covers what is being separated.
Diagram to come
A working steam still in cutaway, reading left to right, drawn large enough to be legible on a phone. Boiler at the left with a steam line into the base of the still body. The still body is cut open to show the charge of plant material resting on a perforated grate, the steam inlet below the grate, and headspace above the charge. Arrows show the steam path going UP THROUGH the charge, not around it. A swan neck off the top carries the mixed vapour to a condenser drawn as a coil inside a water jacket, with cooling water entering at the bottom and leaving at the top so the counter-current is visible. The condenser outlet runs into a separator: a closed vessel with two outlets at different heights, the upper one drawing off a thin floating layer of oil, the lower one running water away continuously. Vapour path and liquid path in two different colours. Mark the temperature at the top of the still as just under one hundred degrees. Every named part carries a label. Someone who has looked at this once should be able to redraw it from memory.
Much of this happens within sight of the field, because cut material starts losing oil at once. The hours between cutting and distilling are part of the oil.
The obvious way to get a volatile compound out of a plant is to heat it until the compound boils off. Limonene boils at around 176 °C, linalool at around 198 °C, beta-caryophyllene at around 265 °C. Plant tissue chars long before that, esters come apart, and what survives rearranges. You would collect an oil, but not the oil that was in the plant.
Steam distillation avoids it. Two liquids that do not dissolve in one another do not behave as a mixture: each pushes its own vapour into the space above as though the other were absent, and the two pressures add. A liquid boils when the pressure it pushes out equals the atmosphere pushing in.
Water alone reaches atmospheric pressure at 100 °C. Inside the still the oil contributes a little pressure of its own, so the mixture boils slightly below 100 °C and sends both components over together. A steam still sits between 99 and 100 °C for the whole run, and beta-caryophyllene, which needs 265 °C alone, leaves at 99 unchanged.
The trick, in one sentence
A still never reaches the boiling point of anything it is distilling. Every compound in the bottle left the plant far below the temperature at which it would have boiled alone — which is why it is still the compound the plant made.
Diagram to come
One vertical temperature scale from 0 to 300 degrees Celsius. Three constituents are marked at their own boiling points: limonene near 176, linalool near 198, beta-caryophyllene near 265. A shaded band across the upper part of the scale marks the region where plant material chars and esters break, and it must visibly overlap all three boiling points. Then a single horizontal line low on the scale, just under 100, labelled as the temperature a steam still actually runs at, with the three constituents drawn crossing that line intact. One thing has to be obvious at a glance: the working line sits far below every boiling point on the scale, and the compounds come over anyway.
The formal statement: in a mixture of immiscible liquids the total vapour pressure is the sum of the pure-component pressures, each one undiminished by the other. That is what immiscibility buys. In an ordinary solution each component’s contribution is cut in proportion to its share; here it contributes everything it would contribute alone.
One wrinkle matters more than it looks. An oil’s constituents are immiscible with water but perfectly miscible with each other, so within the oil phase the proportional rule does apply — each contributes according to its share of that phase. The oil behaves as one immiscible liquid against the water, and as a solution internally. That single fact explains most of what a run does over time.
How much of each ends up in the distillate follows from the pressures and the molecular masses: the mass ratio of two components coming over is the ratio of their partial pressures multiplied by the ratio of their masses. Water is 18; a sesquiterpene is around 204. So a constituent whose vapour pressure at 99 °C is a thousandth of water’s is still carried at something near one part in a hundred by mass rather than one in a thousand — the weight of the molecule buys back most of what its low volatility cost. A great deal of water still has to be boiled to collect a little heavy oil, which is why the heavy end of a run takes hours.
Pressure moves the whole system at once. Under vacuum water boils lower — around 80 °C at half an atmosphere — which is gentler on everything in the charge. But the oil’s vapour pressure falls at that temperature too, so there is less oil per kilogram of steam and the run burns more fuel per gram collected. Above atmospheric pressure the trade reverses: faster, more oil per hour, at a temperature that does more to the molecules. Neither setting is correct in general. Each is correct for a particular material.
Two distillers can take material from the same field on the same morning and produce oils a laboratory can tell apart. Nothing was adulterated; they ran their stills differently. Four things are chosen: pressure, which sets temperature; temperature, which decides what survives; how the charge is packed, since steam takes the easiest path; and duration. A tea tree run is two to three hours, lavender shorter, vetiver and sandalwood days.
Duration matters most, because the oil changes composition while the run is going on. At the start the charge is rich in the lightest constituents, so they dominate the vapour and the first oil to reach the separator is bright, sharp and thin. As they deplete the heavier constituents become the larger part of what arrives, and by the end the distillate is sesquiterpenes and heavy oxygenated compounds. It does not switch; it drifts. Terpenes sets out how size and volatility go together.
Which is why a distiller takes the whole run. The early fraction is the most attractive thing the still will make all day, and there is a standing temptation to stop there and reload. An oil made that way smells wonderful for twenty minutes and has nothing underneath it — much of what people notice in a cheap oil. Purity and quality covers when it is adulteration.
Pressure
Sets the temperature, because the boiling condition depends on it. Most runs are at atmospheric pressure; reduced pressure for material that will not tolerate it, raised pressure where throughput matters more than the top of the profile.
Temperature
Follows from pressure, and this is where distillation is a craft rather than a procedure. Too high and the delicate constituents are changed; too low and compounds that should have come over leave with the spent material.
Duration
The figure belongs to the material — hours for a leaf, days for a dense wood or root, because the molecules held there are heavy and slow and the tissue has to be worked into.
How the charge is packed
Too tight and the material compacts, the steam finds one channel and the rest is barely touched; too loose and it races through with too little contact. Loading a still evenly is a real skill, invisible in the finished bottle.
Esters are the group most at risk inside a still, because an ester sitting in hot water with any acidity present slowly reverts to the alcohol and the acid it was assembled from. Lavender is the case anyone can follow: published profiles put linalyl acetate at 25–45 per cent of the oil and linalool at 20–47 per cent, and linalyl acetate hydrolyses to linalool. So a hotter, longer, higher-pressure run moves the ratio — less ester, more alcohol, a sharper and less rounded oil out of identical flowers.
You cannot read that off a single certificate, because both compounds are genuinely present in the living plant and both ranges are wide. It shows up in a series: the same grower, the same field, several seasons, and a ratio that moves in the year the distillation changed. That is the level at which a serious buyer notices. Reading a constituent profile works through a batch report line by line.
Distillation is usually taught as a separation. But a still is also a reactor: the charge spends hours in heat, water and the mild acidity of the plant’s own sap, and a few molecules arrive changed.
German chamomile has white petals and a yellow centre and nothing blue about it; crush the flowers and your fingers are green. Distil them and what runs into the separator is ink. The compound responsible is chamazulene, and it is not in the plant. The plant holds matricin, a colourless molecule that does not survive a still: it loses a fragment, the remaining skeleton rearranges into a flat fused double ring, and a ring of that shape absorbs yellow and orange light, leaving blue to reach your eye. Blue tansy and yarrow do the same, which is why Deep Blue is the colour it is.
Wintergreen is the same from the other direction: the leaf holds almost no free methyl salicylate, only the aromatic part bound to a sugar, so it is crushed and steeped in warm water for most of a day before the steam is turned on. A constituent report describes the oil, not the plant — a carbon dioxide extract of German chamomile is green, because nothing in it ever got hot enough to turn blue.
The route is decided by what the material tolerates and how much there is to take. Most of this library comes out of a steam still; the exceptions are instructive, and one is not an essential oil at all.
| Method | What it suits | What you get | In this library |
|---|---|---|---|
| Steam distillation | Leaf, flowering top, needle, wood, root, resin | The volatile fraction only, plus a hydrosol | 43 of the 53 singles |
| Cold expression | Citrus peel, and nothing else | Everything the gland held — waxes, pigments, coumarins | All 8 citrus singles |
| COâ‚‚ extraction | Aroma that will not survive a steam run | The closest thing to the intact plant | Madagascar Vanilla |
| Solvent extraction | Flowers that give a still almost nothing | An absolute — not an essential oil | Jasmine, and the osmanthus in Deep Blue |
Citrus explains the rule about temperature. Its oil sits in cavities a fraction of a millimetre under the coloured surface of the rind, so the fruit is scored and pressed and a centrifuge separates oil from juice. No heat at any stage. Steam-distil a lemon rind and you get an oil, but not lemon: flatter, drier, faintly like turpentine, because the molecules carrying that character are the lightest and most reactive in the bottle.
Pressing has one consequence steam does not. Steam carries only what will evaporate; pressing carries everything the cavity held, including heavy, barely volatile material — waxes, pigments, and a group of compounds called coumarins. Those react with ultraviolet light on skin. After applying any expressed citrus oil to skin, always diluted in a carrier oil, keep that skin out of direct sunlight and away from tanning beds for up to 12 hours.
Above about 31 °C and 74 atmospheres, carbon dioxide becomes a supercritical fluid — it flows like a gas and dissolves like a liquid. Run it through plant material and it carries the aromatic compounds out; drop the pressure and it leaves as a gas, no residue. The temperature never goes far above body heat, so what comes out is as close to the intact plant as any method gets — and is not interchangeable with the distilled oil of the same plant. Which products are available differs from country to country.
An absolute is not an essential oil
Jasmine and osmanthus give a still almost nothing — the blossom will not survive steam and does not hold enough to be worth the fuel. They are extracted with a solvent, which is then removed, leaving a thick, waxy, deeply coloured material called an absolute. It was not distilled and it is not an essential oil. Jasmine is listed as an absolute.
A separator produces two liquids and only one is the product. The other is condensed steam — a hydrosol, of which rose water is the oldest example.
A hydrosol is not dilute essential oil. It is a different composition — the water-soluble corner of the same plant, with the greasy majority missing.
Most of what is in an oil will not dissolve in water at all — the hydrocarbons, the bulk of the bottle, are excluded almost completely. What the water keeps is the minority of constituents polar enough to be comfortable in it, plus a saturated trace of the oil itself, usually well under a tenth of one per cent.
Water molecules hold on to each other through hydrogen bonds, and they will admit a guest that can join in. A hydrocarbon terpene cannot — no polar group, nothing to bond with — so the water closes ranks and the excluded molecules gather with their own kind. That exclusion is the layer in the separator. Attach an oxygen-bearing group to the same skeleton and it changes: an alcohol can hydrogen-bond, and small alcohols such as linalool have a real, measurable solubility in water.
Rose is where this stops being theoretical. Much of rose’s character is carried by phenylethanol, which is soluble enough that a great deal of it goes into the water rather than the oil — which is why rose water smells as much of rose as it does, and part of why rose oil is as scarce as it is. Distillers answer with cohobation: the hydrosol from one run is returned to the still and re-distilled with the next charge, so the dissolved fraction is recovered instead of poured away. It is standard for very few materials, and it exists because of one constituent’s solubility.
It is mostly water with nothing preserving it, so left warm it grows things; cold and sealed it keeps for months rather than years.
Yields differ by orders of magnitude, and the figures explain most of what puzzles people about price.
Lavender
Under one per cent of harvest weight — roughly 150 pounds of tops per pound of oil, two kilograms per 15 mL bottle.
Peppermint
A third to a half of one per cent of fresh weight — two to four kilograms a bottle.
Tea tree
One to two per cent of fresh leaf and branchlet.
Oregano
About one per cent of dried leaf and flower — a kilo and a half dried, five to seven fresh.
Rose
Approximately 60,000 roses for one ounce of oil, sold in a 5 mL bottle — a size that is not a marketing decision.
Photograph to come
A single 15 mL amber bottle standing on a plain surface beside the quantity of harvested material it took — for lavender, roughly two kilograms of cut flowering tops, which is a genuine armful and must look like one. Shot from slightly above with both the pile and the bottle in focus. Nothing styled: no linen, no props, no sprigs arranged into a composition. The entire value of the photograph is the ratio, so the bottle has to read as small without being made small by the lens.
The spread comes from where a plant keeps its aromatics. A leaf or rind stores oil in bulk; a flower releases its volatiles into the air rather than stockpiling them, so at any moment there is very little to take. Plants and their oils covers what a plant makes them for.
Yield sets most of the price; the rest is how long the still is occupied and how much is done by hand. When an oil comes in a 5 mL bottle rather than 15 mL, the bottle is telling you what the field cost. Sourcing is the ground behind that.
Keep oils away from eyes and inner ears, and out of reach of children. If you are pregnant, nursing, under medical care or taking medication, speak to your healthcare practitioner before use. Nothing here is intended to diagnose, treat, cure or prevent any disease.
Lesson 3 of 21 · Essential Oils