AromatherapyEssential oils — the real effects, and the real toxicology
Smell reaches the brain by a shorter route than any other sense, and that part is genuinely remarkable. Essential oils are also the most concentrated plant products most people will ever handle — and are sold as if they were mild.
⚠ The three rules, before anything else
Never swallow essential oils. Not a drop in water, not in a capsule, not “therapeutic grade.” There is no regulatory grade, and lavender oil alone is documented as causing severe sedation in children at very small amounts.
Never apply them undiluted. A drop of essential oil is the concentrated extract of a great deal of plant.
Cats and birds should not be in a room you are diffusing in. Birds have exceptionally vulnerable airways, and cats process these compounds poorly.
Everything else on this page is detail. Those three prevent almost all the harm.
How to read this page
Part I is the practice and how to do it safely, including actual dilution numbers. Part II is what the evidence shows, what is oversold, and the industry that did the overselling.
What is actually in the bottle
An essential oil is not oil. It is the volatile aromatic fraction of a plant, usually steam-distilled or cold-pressed, concentrated many times over.
The concentration is the thing people underestimate. Producing a small bottle of rose oil takes an extraordinary quantity of petals; a bottle of lavender takes a field. You are not handling a herb. You are handling a chemistry set derived from one.
Attested The word is modern. “Aromathérapie” was coined by the French chemist René-Maurice Gattefossé in the 1930s. The familiar story — that he burned his hand, plunged it into lavender oil and was astonished at the healing — is his own account, and it is the founding anecdote of an entire industry. Aromatic plants are ancient; distilled essential oils used this way are a twentieth-century practice.
There is no regulatory body that certifies essential oil grades. “Therapeutic grade,” “clinical grade” and “CPTG” are marketing terms coined by the companies that use them, not standards anybody audits.
What actually tells you something: the Latin botanical name, the country of origin, the extraction method, the plant part, and a batch GC/MS report. A company that gives you those is telling you more than one that gives you an adjective.
Notes — how a blend is actually built
Perfumery borrowed the language of music, and it is genuinely useful. Every oil evaporates at its own rate, and a blend that only uses one speed smells thin and disappears fast.
Bright, sharp, the first thing you meet. Citrus, peppermint, eucalyptus, lemongrass. They lift and they leave.
The body of the blend and usually most of it. Lavender, geranium, chamomile, rosemary, marjoram, tea tree. This is where the character lives.
Deep, heavy, slow. Frankincense, sandalwood, vetiver, patchouli, cedarwood, myrrh. They anchor everything above them and stop a blend evaporating in twenty minutes.
Roughly 30% top, 50% middle, 20% base. In practice, for a ten-drop blend: 3 top, 5 middle, 2 base.
That is a starting point and not a law. Base notes are powerful and cheap to overdo — one drop of vetiver will announce itself over nine drops of anything else. When in doubt, use less base.
Blend the essential oils together first, in a small bottle, before adding carrier. Let it sit a day or two if you can — blends genuinely settle and round out. Then dilute to the percentage you need.
Building a blend for a purpose
Pick what you actually want, and this gives you a working recipe with real drop counts. Every one of these is a starting formula, not a prescription — substitute freely for what you have and what you like, because a scent you dislike will not relax you however correct it is on paper.
The oil reference
Thirty oils — what each is traditionally used for, which note it sits in, how to use it, and what to watch. Filters dim rather than hide, and the safety line never dims.
Purpose
Note
Six things to actually make
Cheap, quick, and each one teaches a method you can then vary endlessly.
15–20 drops of neat blend onto the cotton wick, then cap it. Lasts months. The most controllable method there is — you get the dose, nobody else in the room does, and it fits in a pocket.
A 10ml roller: 4 drops total for 2%, or 2 drops for 1% if it is going near the face. Top with carrier. Pulse points, temples, back of the neck.
100ml bottle: 25–30 drops, a tablespoon of vodka or witch hazel to disperse, then water. Shake before every use — without the alcohol it will separate and spray unevenly.
5–8 drops emulsified first in a tablespoon of carrier oil, full-fat milk or unscented liquid soap. Never drop neat oil into water — see §07.
30ml carrier, 12 drops for 2%. Warm it in your hands first. This is the classic method and the reason most of the anxiety evidence exists.
Hot water, 2–3 drops only, towel over the head, eyes closed. Very effective for congestion and very easy to overdo. Not for young children, and not with peppermint or eucalyptus near their faces.
Nearly everything in aromatherapy is one of these six methods with different oils in it. Learn the methods and you stop needing recipes.
Dilution, with the actual numbers
The single most useful practical section on this page. Carrier oil is not a formality — it is what makes the practice safe, and it also helps the oil spread and absorb.
Carriers: jojoba, sweet almond, fractionated coconut, grapeseed, olive. Any of them. The carrier is doing a physical job, not a magical one.
A diluted drop on the inner forearm. Wait 24 hours, not five minutes — sensitisation reactions are frequently delayed, and phototoxic reactions can peak 36 to 72 hours after sun exposure.
And sensitisation is permanent. Once you have reacted to an oil, you generally react to it for life — which is why over-use of a favourite oil is the most common way people lose it.
How it is actually used
- Diffusion. Water diffusers, nebulisers, or a few drops on a tissue. Intermittent is better than continuous — 30 to 60 minutes on, then off. A saturated room is not more effective, only more irritating.
- Direct inhalation. A drop on a cotton pad or an inhaler stick. The most controllable method and the one with the least that can go wrong.
- Diluted topical. In a carrier, for massage or a specific area. See §02 for the numbers.
- ⚠ Baths — and this is the one people get wrong. Oil does not disperse in water. Dropping neat oil into a bath leaves it floating in concentrated beads that then meet your most delicate skin. Emulsify first in a tablespoon of carrier oil, full-fat milk, or unscented liquid soap.
- ⚠ Never internally. Not in water, tea, capsules or honey. This is covered again in §06 because it is where the serious poisonings come from.
Why smell is genuinely different
Studied Olfaction takes a shorter route than any other sense. Sight, sound and touch are relayed through the thalamus before reaching the cortex. Smell projects directly to the olfactory bulb and from there to the amygdala and hippocampus — structures central to emotion and memory.
That is why a smell can produce a feeling before you have identified it, and why scent-triggered memory arrives with unusual emotional force. This part is real neuroanatomy, not a metaphor, and it is the honest foundation of the whole practice.
What it does not establish is that any particular oil has any particular therapeutic property. Those are separate claims, and they need separate evidence.
What the evidence supports
Anxiety and mood Studied
The best-supported claim. Lavender in particular has repeated trials showing modest reductions in anxiety, especially in stressful settings — pre-operative waiting, dental clinics, intensive care. Effects are real and small, and studies are frequently hard to blind, since participants can obviously tell whether the room smells of lavender.
Sleep Studied
Modest evidence for improved subjective sleep quality. Subjective is doing work in that sentence — people report sleeping better more reliably than instruments record it.
Nausea, headache, alertness Studied
Peppermint has some support for nausea and for tension headache applied topically diluted. Alerting effects of citrus and peppermint are among the more consistent findings.
Antimicrobial activity Contested
Many essential oils are genuinely antimicrobial in a petri dish. Tea tree in particular. ⚠ That does not transfer to treating an infection in a person, and it is the single most over-extrapolated finding in this field.
And the honest confound
Expectation and pleasantness explain a great deal. A smell you find lovely, in a calm room, while somebody attends to you, will reduce anxiety. That is a real effect with an ordinary explanation — and it is not diminished by being explicable.
⚠ The toxicology
Contested In 2007, Henley and colleagues reported three boys with prepubertal gynecomastia that resolved when lavender and tea tree products were stopped, and showed both oils had oestrogenic and anti-androgenic activity in vitro. A 2019 report added further cases.
The rebuttals are substantial and this page will not hide them. A review applying Hill's criteria concluded the association rests on around twelve patients across five publications, that dermal penetration is likely insufficient for the exposures described, that both conditions often resolve spontaneously anyway, and that a causal link has not been shown. A separate critique argued the in-vitro results may reflect contamination from plastic labware rather than the oils themselves.
So the honest position: a real signal, a plausible mechanism, and evidence too weak to call it established. Worth knowing if you are applying lavender products daily to a child. Not worth panic, and not worth dismissal either.
⚠ The industry
The steelman
- The neuroanatomy is genuinely unusual. No other sense has such direct access to emotion and memory, and building a practice on that is not naive.
- The anxiety evidence is real, even if modest — which puts aromatherapy ahead of most of what this library examines.
- It is cheap, pleasant, and easy to do correctly. Diluted, inhaled, intermittent: very low risk and a genuine improvement to a room and a mood.
- Scent marks transitions. The same oil for the same ritual becomes a conditioned cue, and that mechanism — ordinary classical conditioning — is powerful and entirely yours to build.
- And the plants are real. Aspirin came from willow; a great many drugs began as plant chemistry. The scepticism here is about specific claims, never about the idea that plants contain active compounds.
What a new seeker should actually ask
- ⚠ Is anybody telling me to swallow this? Then stop listening to them about oils.
- ⚠ Is it diluted? If it goes on skin, it should be. §02 has the numbers.
- ⚠ Is it citrus, and am I going outside? Cold-pressed bergamot, lime, lemon, grapefruit and bitter orange can burn. Steam-distilled or FCF versions do not.
- ⚠ Is there a cat, a bird, a baby or an asthmatic in this room? Ask before you diffuse.
- Does the seller earn more when I buy more?
- Can they tell me the Latin name, origin and extraction method? If they can only tell me the grade, that is an adjective, not information.
- Is this instead of treatment for an infection? Antimicrobial in a dish is not antimicrobial in a person.
Provenance
Phototoxicity Studied
Furanocoumarins (psoralens), principally bergapten and oxypeucedanin, in expressed citrus peel oils — Citrus bergamia, C. aurantium, C. limon, C. aurantifolia — and in angelica root and tagetes. Not present in distilled citrus peel oils. Reactions from pigmentation through blistering to full-thickness burns; onset within 24 hours, peak 36–72 hours post-UVA. IFRA maximum for bergamot in leave-on products: 0.4%. Risk considered removed if skin is protected from UVA for at least two hours. Historically termed berloque or bergapten dermatitis; Klaber coined “phytophotodermatitis.”
The endocrine question Contested
Henley, Lipson, Korach & Bloch (2007), Prepubertal gynecomastia linked to lavender and tea tree oils, NEJM — three cases plus in-vitro oestrogenic and anti-androgenic activity. Ramsey et al. (2019) added premature thelarche cases. Against: a review applying Hill's criteria across five publications and roughly twelve patients concluded dermal penetration is likely insufficient, both conditions frequently regress spontaneously, and causation has not been demonstrated; a published letter argued the in-vitro findings may reflect endocrine-disrupting contamination from plastic labware.
Ingestion toxicity Attested
US Poison Control guidance on lavender oil: toxic if swallowed; very small amounts toxic in children, causing severe sedation; larger amounts producing ataxia, headache, confusion, lethargy, coma, blurred vision and respiratory difficulty. Wintergreen (methyl salicylate), eucalyptus and camphor, pennyroyal (pulegone hepatotoxicity), tea tree and clove are the oils most associated with serious ingestion cases.
Gattefossé Attested
René-Maurice Gattefossé coined “aromathérapie” in the 1930s. The burned-hand account is his own, and this page reports it as an origin story rather than as a demonstrated result.
Where the record is thin Contested
Dilution percentages here follow widely-taught aromatherapy guidance rather than a regulatory standard, and practitioner bodies differ on the exact figures for children. Pet toxicity is well attested clinically but the veterinary literature is thinner than the strength of the warning implies — the warning is retained because the consequence is severe and the cost of caution is nil.
✦ For practitioners
- ⚠ Never suggest ingestion. Ever. Whatever your training said, whatever the company says. This is the line.
- Ask the room before you diffuse. Asthma, pregnancy, cats, birds, babies. Ten seconds, and it prevents most of the harm on this page.
- Teach dilution numerically, not by feel. “A few drops” is how people end up at 10%.
- Flag citrus and sun in the same sentence, every time. Phototoxic burns are the injury most likely to come out of a well-meant massage.
- Tell people the pharmacy bottle is fine. If your practice depends on selling a proprietary grade, it is a sales relationship, and your client should be told which one they are in.
✦ For those guiding others
- Ask where the oils came from and who advised them. If the answer is an upline rather than a shop, the ingestion advice is probably close behind.
- Watch for oils replacing treatment — especially for infections, where the petri-dish evidence is most persuasive and least applicable.
- ⚠ Households with small children. The bottles look harmless and smell like sweets. Storage is the whole conversation.
- Do not sneer at the scent itself. The neuroanatomy is real, and a person who finds a smell calming is not being fooled.
We are not vouching for these, only pointing at them.
- Tisserand & Young, Essential Oil Safety — the standard reference, and unusually rigorous for this field.
- Your national poison centre's page on essential oils. Free, sober, and specific about doses.
- The IFRA standards for the actual concentration limits behind §06.
- The Henley 2007 paper and the Hill's-criteria review together. Read as a pair, they are a good short lesson in how a real signal and an unproven cause differ.