Non Toxic Essential Oils (2026): What "100% Pure" Is Allowed To Mean
The 30 Second Summary
- There is no one percent law. The widely repeated claim that a bottle can be sold as 100% pure if only one percent of it is real oil does not appear in any FDA regulation. The actual position is less comforting: no federal rule defines pure, natural or therapeutic grade for an essential oil, and none requires a composition test before those words go on a label. There is no threshold to hold anyone to because there is no definition.
- Pure is not the same as safe. Pulegone, estragole, methyleugenol and safrole are on California's Proposition 65 carcinogen list, and all four occur naturally in authentic oils: pulegone in peppermint, estragole in basil and fennel, methyleugenol in basil and nutmeg. The entire industry argument is fought over authenticity, which is the wrong axis.
- One document settles the authenticity question, and it needs two things. A composition report from a laboratory that does not work for the brand, carrying the lot code printed on your bottle. A report keyed to a barcode describes the product line, not your bottle. In house testing is a company marking its own homework. "GC/MS tested" on a label gives you neither.
- The umbrella to watch sits on blends, not single oils. A single oil named by botanical species is already a complete ingredient list. "Fragrance", "parfum", "essential oil blend", "natural fragrance" and a mood name with nothing underneath it all name nothing you can check.
- If fragrance makes you ill, two real disclosure tools exist. California's fragrance right to know database is public and searchable, and any cosmetic newly placed on the EU market since 31 July 2026 must name 81 allergens instead of 26. Neither depends on trusting a brand.
- Keep oils in glass. D-limonene is 60 to 97 percent of citrus oils and is sold industrially as a solvent for dissolving polystyrene. Never decant into plastic, and never drip oils into a plastic diffuser or humidifier tank.
- Checkable down to your own bottle: Plant Therapy Lavender, the one brand here that publishes a separate outside laboratory report for each production lot, keyed to the code on the bottle.
- Organic, same reports: Plant Therapy Organic Lavender, USDA certified organic with the same lot keyed report published beside it.
- Blends with every oil named: Aura Cacia Peppermint and its mood named blends, which list every constituent oil by botanical name. Its own testing is in house and its published reports are keyed to the barcode rather than your lot.
A reader with multiple chemical sensitivity wrote to ask about a specific piece of labelling law: the rule that supposedly lets a bottle be sold as 100% pure when only one percent of its volume is the real thing. She has cut her own use of essential oils almost to nothing, not because she reacts to them, but because she does not want to recommend something to other people that might turn out to be mostly synthetic. That instinct is sound, and the law she was told about turns out not to exist. What exists instead is a bigger hole in the same wall.
The one percent rule is not a rule
No FDA regulation permits a product that is one percent essential oil to be labelled 100% pure. The same claim circulates with five percent and with 51 percent in place of one, which is usually a sign that a number has been passed along rather than read. There is no such provision in the cosmetic labelling rules, the food additive rules or the drug rules, and the FDA has never issued a determination along those lines.
The reason the story is so persuasive is that something adjacent to it is true. The cosmetic ingredient rules do contain a one percent threshold, at 21 CFR 701.3(f). It says that ingredients present at one percent or less may be listed in any order, after the ones present at more than one percent, instead of strictly by descending weight. That is a rule about the order of a list. It has nothing to do with what may be called pure, and it does not let anything be left off the list.
The absence of the rule is worse news than the rule would have been. A one percent threshold would at least be a number, and a number can be measured and enforced. What actually governs the word pure on an essential oil is nothing at all.
Here is the position in full. "Pure", "natural" and "therapeutic grade" have no legal definition for an essential oil anywhere in United States law. No agency certifies them. No test is required before they are printed. The FDA's authority over a bottle of essential oil depends on what the seller claims it does: described as a scent it is a cosmetic, described as a treatment for a disease it becomes an unapproved drug, and it is that second line, rather than any question of composition, that the agency has actually enforced. In September 2014 it sent warning letters to the two largest sellers in the category over disease claims their consultants were making, including claims about Ebola.
Where the phrase "therapeutic grade" appears with initials after it, those initials are typically the seller's own registered trademark. That is worth understanding precisely, because it is the exact inverse of how a certification reads. A trademark means the company owns the words and can stop anyone else using them. It does not mean anybody outside the company checks the thing the words describe.
What the umbrella terms are allowed to hide
The reader also asked about proprietary blend labelling, and the two umbrellas work differently. The weaker of the two, by a wide margin, is the cosmetic one. Under 21 CFR 701.3(a), every ingredient in a cosmetic must be declared in descending order of predominance, with one carve out: fragrance and flavour may be declared simply as "fragrance" and "flavor". Where an ingredient has been accepted as a trade secret under 21 CFR 720.8, the label may finish with the phrase "and other ingredients". A single word can therefore stand in for a mixture of fifty, a hundred or more individual compounds, and the regulation's own rationale is that fragrance formulas are the cosmetic components most likely to qualify as trade secrets.
Supplement labelling is stricter, which surprises people who assume the reverse. Under 21 CFR 101.36, a proprietary blend must carry the total weight of the blend and must name every ingredient in it, in descending order of predominance. What it may withhold is the amount of each one. So a proprietary blend tells you what is in there and not how much; a fragrance tells you neither. If you are triaging labels by how much they conceal, the fragrance umbrella is the one to treat as opaque.
A mood name on the front of a bottle is not the problem. The question is whether the oils underneath it are written down anywhere you can read them.
This matters less than people expect on a bottle of single oil and more than they expect on everything else. A bottle labelled "Lavandula angustifolia" is already a complete ingredient list, because there is one ingredient. The gap opens on blends, roll ons, diffuser pods and every scented product in the house, where a mood name on the front is doing the work that an ingredient list should be doing. The useful discovery while checking this was how cleanly some makers handle it: Aura Cacia's blends carry names like Fresh, and the ingredient list underneath reads out every oil by botanical species, Balsam Peru and tea tree and peppermint included. A mood name with the oils published under it is fine. A mood name with nothing under it is the same failure as "fragrance".
Pure is not the same as safe
The whole essential oil argument is conducted on one axis. Brands accuse each other of dilution and adulteration, buyers learn to spot synthetic linalool, and the implicit promise is that an oil which is genuinely what it says it is must therefore be fine. That promise does not hold, and California's own hazard list is the simplest way to see why.
Four compounds on the Proposition 65 carcinogen list occur naturally in ordinary essential oils. Safrole was listed in 1988, estragole in 1999, methyleugenol in 2001 and pulegone in 2014. Pulegone is a normal constituent of peppermint, catnip and pennyroyal oils. Estragole occurs in basil, fennel, anise and tarragon. Methyleugenol occurs in basil, nutmeg and allspice. None of these is an adulterant. They are what the plant makes, and a perfectly authentic, correctly labelled, outside verified oil can be full of them.
Verification answers the columns. It says nothing about the rows, which is why a brand can win the purity argument outright and still sell you a listed carcinogen.
California's enforcement record carries the point. Searching the Attorney General's 60 day notice register for essential oil products returns twelve notices. The oldest, from 2001, alleged estragole in aromatherapy essential oils. Three more allege DEHP, a phthalate, in essential oil gift kits and cases rather than in the oils themselves. And the most recent was filed on 21 October 2026, naming pulegone in a line of essential oil roll ons sold by one of the largest brands in the category. A notice is an allegation rather than a finding, and it caps how far we will go in either direction. But the chemistry behind it is not in dispute, and it is not an adulteration story.
The one test that works, and its two halves
Gas chromatography with mass spectrometry separates an oil into its constituent compounds and names them. Run against the ISO standard for that species, it is a genuinely good tool: it catches dilution with a cheaper oil, and chiral analysis of the ratio between mirror image forms of linalool will catch synthetic linalool added to lavender even when the overall profile looks normal. Published case studies on lavender, lemon and tea tree have found exactly these adulterations in commercial samples.
So a composition report is the right document to ask for. The trouble is that almost every brand in the category now says the words "GC/MS tested" somewhere, and the phrase is doing no work. Both of the bottles recommended at the bottom of this article say something like it on their Amazon listings. What separates a report you can use from a claim you cannot is two things, and both have to be true.
Was it run by a laboratory that does not work for the brand, and does it carry the lot code printed on the bottle in your hand? A report keyed to a barcode describes the product line. A report run in house is the company checking its own work.
Checking the two largest transparency claims in the category against those questions produced a clear split. Plant Therapy's product pages carry a Test Reports tab holding separate downloadable reports for each production lot, named by lot code, and the company states the testing is done by an outside laboratory. On the lavender page at the time of writing there were three, for lots L40122R, L40123R and L40124R. Match the code on your bottle to the right file and you are reading a measurement of what you actually bought. That is the strongest form this evidence takes.
Aura Cacia publishes a table of reports too, and the difference is instructive. Its reports are indexed by item number and barcode rather than by lot, so the file you download describes that product generally rather than the bottle on your shelf, and the company states the oils are tested in its own laboratories in Iowa rather than by an outside lab. That is more than most of the category offers and considerably less than a lot keyed outside report. Both facts belong on the page.
What a composition report does not tell you
This is the part that gets lost, and it is the reason nothing in this category reaches the top of our scale on testing. A GC/MS report is a composition document. It answers the question "is this the oil it claims to be", and it is not a contaminant screen. It is not run for lead, cadmium or arsenic. It is not run for pesticide residue. It is not run for phthalates. A brand can publish a flawless lot keyed report and have no idea whether there is a heavy metal in the bottle, because nobody asked that question of it.
We searched the independent testers whose results our verdicts normally rest on, and none of them has published a contaminant panel on essential oils. There is no Lead Safe Mama report, no Mamavation panel, no Consumer Reports table. That is a gap rather than a finding, and we do not mark a product down for a test nobody has run. But it means the honest status of the whole category on independent testing is "nobody has looked", and anyone telling you an essential oil has been proven clean is describing a document that does not exist.
The two disclosure levers that actually exist
For anyone whose problem with fragrance is immediate and physical rather than theoretical, the useful news is that two mechanisms now exist which do not require trusting a brand's word about anything.
The first is Californian. The Cosmetic Fragrance and Flavor Ingredient Right to Know Act of 2020 was signed on 30 September 2020 and became enforceable on 1 January 2022. It requires any company selling retail cosmetics or professional salon products in California to report the presence of any fragrance or flavour ingredient appearing on one or more of 23 authoritative hazard lists to the California Safe Cosmetics Program, which publishes it in a searchable public database at cscpsearch.cdph.ca.gov. It was the first law in the United States to require fragrance ingredient disclosure, and because the reporting duty attaches to selling in California it covers a very large share of what is on American shelves. It will not list a benign fragrance component, since only hazard listed ones are reportable, which for the purpose of avoiding a reaction is the right half of the list to have.
The second is European, and it changed three months ago. Commission Regulation (EU) 2023/1545 expanded the list of fragrance allergens that must be named on a cosmetic label from 26 to 81, following the identification of 57 further allergens by the Scientific Committee on Consumer Safety. Labelling is required once a listed allergen exceeds 0.001 percent in a leave on product or 0.01 percent in a rinse off one. Anything newly placed on the EU market has had to comply since 31 July 2026, and stock already on the market has until 31 July 2028. In practice this means the EU labelled version of a product frequently discloses considerably more than the American one, and a pack bought in or imported from Europe is worth reading for that reason alone.
Set against those two, the federal position is thin. The Modernization of Cosmetics Regulation Act of 2022 instructed the FDA to issue a proposed rule on fragrance allergen labelling by 29 June 2024. That deadline passed without a proposal. The regulatory agenda published in July 2026 now targets November 2026 for the proposal, and Congress directed the agency to look at international and state requirements, the EU's included, when writing it. A proposed rule is the beginning of a process, not a label change, so nothing on an American shelf has moved yet.
On EWG, precisely
The reader's note said that EWG ignores fragrance and perfume ingredients, and that indoor air quality specialists she had met did not know this. The concern underneath that is correct and the mechanism is slightly different, which is worth getting right because the distinction changes what the score is good for.
Skin Deep does not ignore fragrance. It carries "fragrance" as an entry and scores it 8 out of 10, near the top of its hazard range. What it cannot do is assess the mixture, because the mixture is not disclosed to it: Skin Deep's scores are built from publicly available ingredient lists, so where a label says "fragrance" the database has a single generic hazard score standing in for several dozen compounds nobody outside the manufacturer has seen. Every entry also carries a second score for data availability, and that is the half that tells you how much is actually known. EWG Verified, which is a different programme, does require a manufacturer to disclose every fragrance component directly to EWG as a condition of the mark.
So the practical upshot matches the reader's instinct without matching her description. A product can show a reassuring overall score while the single largest unknown in it was never evaluated, because a generic placeholder is not an assessment. If you react to fragrance, an ingredient score database is the wrong instrument, and the California database is the right one, because it is built on mandatory disclosure rather than on reading labels.
The part that belongs to us: oils are solvents
Essential oils are the rare consumer product that attacks its own container, and the mechanism is not obscure. D-limonene makes up roughly 60 to 97 percent of cold pressed citrus oils, and the same compound is sold industrially as a solvent, including for dissolving polystyrene. Terpenes are small non polar molecules that slip between polymer chains and push them apart, so neat oil will craze and dissolve polystyrene on contact, swell and distort PET, and stress crack thinner polyethylene until it weeps.
Three practical consequences follow. Keep oils in glass, which is why the category sells amber bottles and why the only plastic part is a small orifice reducer chosen to resist them. Never decant an oil into a plastic bottle, however convenient, and do not mix a blend in a plastic measuring cup. And do not put oils in an ultrasonic diffuser or humidifier with a plastic tank, which combines the solvent problem with a worse one: that kind of machine aerosolises whatever is in the tank and blows it into the room for you to breathe, which is the argument against scenting a room that way regardless of what the tank is made of.
If you have small children and own wintergreen or birch oil, check the cap. Both are largely methyl salicylate, which United States law requires to be sold in child resistant packaging under the Poison Prevention Packaging Act, and a steady run of recalls since 2019 has been for failing that requirement rather than for anything wrong with the oil. One recall in April 2021 covered about 1.3 million bottles. Recalls on the same ground have also covered wintergreen oils from Aromatics International, Pranarom, Ananda Apothecary, Best Nutritionals and, as recently as January 2025, an Amazon seller's listing.
The one case where the hazard was not the chemistry
It is worth knowing the single worst documented outcome in this category, because nothing about it would have been caught by any purity test. In 2021 four people in Georgia, Kansas, Minnesota and Texas developed melioidosis, a rare and frequently fatal infection caused by the bacterium Burkholderia pseudomallei. Two of them died. Investigators found the organism in a bottle of Better Homes & Gardens Essential Oil Infused Aromatherapy Room Spray with Gemstones taken from one of the patients' homes, and about 3,900 bottles across six scents were recalled from Walmart. The product was made in India and had been on sale from February to October 2021. The oils in it were beside the point. A spray bottle of water, botanical material and decorative stones is a growth medium, and nobody had tested it as one.
How we vet an essential oil
Five fronts, in order, and the first two are the ones that produce verdicts here. The label's ingredient list, read from the maker's own product page, with a single oil named by botanical species counting as complete and an umbrella term counting as a disclosure failure. The contact materials, which for this category means glass with a small plastic dispenser, and a plastic dispenser on a glass bottle is noted on the card rather than counted against it. Independent tests, which in this category means a dated search that came back empty. Lawsuits, read from California's 60 day notice register rather than from summaries of it. And recalls, read from the CPSC's own pages.
Before anything is rated good it also gets a pass whose only purpose is to disqualify it, searching the brand against lawsuit, class action, recall, CPSC complaint, FDA warning letter and attorney investigating. That pass is what produced the wintergreen recall list above, including recalls against brands otherwise well regarded for their transparency, and it is why a brand's reputation for publishing reports is never taken as a substitute for checking the register.
The picks
One caveat belongs above these rather than below them. The honest first recommendation for anyone whose reaction to fragrance is physical is the unscented version of whatever you are buying, because it is the only version whose scent chemistry is not a question at all. What follows is for the case where the scent is the point, which is most people most of the time, and where the useful question becomes which bottle you can actually verify.
Plant Therapy Lavender
Amber glass, steam distilled Lavandula angustifolia, botanical name and country of origin stated. A separate outside laboratory composition report is published for each production lot, keyed to the code on the bottle.
Plant Therapy Organic Lavender
Amber glass, USDA certified organic Lavandula angustifolia, single species named on the label. Same lot keyed outside laboratory reports published on the product page beside it.
Aura Cacia Peppermint
Amber glass, Mentha piperita named on the label, and the mood named blends list every constituent oil by botanical species. Widely stocked rather than direct order only.
careful: testing is in house and the published reports are keyed to the barcode, not to your lot.
Every brand we checked
Five sellers, read on their own product pages for composition and materials, on California's 60 day notice register for litigation, and on the CPSC's recall pages for recalls, all on 26 October 2026. No independent contaminant testing exists for any of them, which is why no row claims one.
| Brand | Composition you can read | Legal record | Verdict |
|---|---|---|---|
| Plant Therapy single oils | Outside laboratory report per lot, keyed to the bottle code | Checked, nothing found | The only lot keyed outside report here |
| Aura Cacia single oils and blends | Every oil named by species, blends included; reports keyed to barcode | Checked, nothing found | In house testing, named composition |
| Edens Garden | States a report per batch; the lot keyed chain could not be opened and confirmed | Checked, nothing found | Returning to it |
| doTERRA | Single species named; "CPTG" is the company's own trademark, not a standard | Prop 65 lead notice on its supplements, filed 2022, open; 1.3 million bottle recall, 2021 | Open allegation, dated recall |
| Young Living | Single species named; "Seed to Seal" is the company's own programme | Prop 65 pulegone notice on roll ons filed 21 October 2026; federal convictions 2017 | Allegation five days old |
The ones we did not pick
doTERRA
A Proposition 65 notice alleging lead in its dietary supplements was filed on 3 February 2022 by the California Toxic Research Institute and remains open on the register, with no settlement or judgment recorded. Separately, on 1 April 2021 the company recalled about 1.3 million bottles of three roll on products for holding methyl salicylate without child resistant packaging, recall 21734, with no injuries reported. That recall is old enough that it informs the picture rather than sets the verdict; the open notice is what holds this at careful.
The letters CPTG, which the brand's marketing builds on, are its own registered trademark rather than a grade defined or audited by anyone outside the company. That is not a finding against the oils. It is a reason not to read the phrase as a certification.
Young Living
On 21 October 2026, five days before this was written, the Center for Consumer Safety filed a Proposition 65 notice naming pulegone in essential oil roll ons sold by Young Living, with Amazon named alongside. It is open and unresolved. An allegation is not a finding and caps a verdict at careful rather than producing a skip, and in this instance the chemistry is unsurprising: pulegone is a natural constituent of mint oils, which is the whole argument of this article rather than an accusation of adulteration.
The company also has a federal criminal record on sourcing. On 18 September 2017 it pleaded guilty in the District of Utah to misdemeanour violations of the Lacey Act and the Endangered Species Act over trafficking in illegally harvested Peruvian rosewood and in spikenard oil exported from Nepal without a CITES permit, conduct running from June 2010 to October 2014. It paid 500,000 dollars in fines plus 135,000 in restitution and a 125,000 community service payment, and served five years of probation with a compliance plan. The company had disclosed the violations itself in 2015 and cooperated. It is nine years old and it is about provenance rather than what is in the bottle, which is why it is recorded here and not treated as current.
Frequently asked questions
Can a bottle be labelled 100% pure if only one percent of it is essential oil?
No federal rule says that, and the claim does not appear anywhere in FDA regulation. The real situation is less reassuring. No federal rule defines pure, natural or therapeutic grade for an essential oil, and none requires a composition test before those words go on a bottle. There is no threshold to point at because there is no definition to hold anyone to. The one percent figure in the cosmetic rules is 21 CFR 701.3(f), which lets ingredients present at one percent or less be listed in any order rather than strictly by weight. It governs the order of a list, not the contents of a bottle.
What does therapeutic grade mean on an essential oil?
Nothing standardised. No government body or independent standards organisation defines or certifies a therapeutic grade for essential oils. Where the phrase appears with initials after it, it is usually the seller's own registered trademark, which means the company owns the name rather than that anyone outside the company audits it. Judge the bottle on its published composition report instead.
Is fragrance worse than essential oil blend on a label?
They fail in the same way. Fragrance and parfum are legal umbrellas under 21 CFR 701.3(a) that can stand for dozens of undisclosed components. Essential oil blend, natural fragrance and plant based oils are not legal terms at all, and they name just as little. What clears the test is every oil named, ideally by botanical species. A mood name on the front of the bottle is fine as long as the oils are listed somewhere you can read them.
Are pure essential oils safe?
Purity and safety are different questions, and the marketing argument is almost entirely about the first one. Pulegone, estragole, methyleugenol and safrole all sit on California's Proposition 65 carcinogen list and all occur naturally in authentic, correctly labelled essential oils: pulegone in peppermint and pennyroyal, estragole in basil, fennel and anise, methyleugenol in basil, nutmeg and allspice. A perfectly authentic oil can carry them. Verifying authenticity does not address them.
How do I check an essential oil GC/MS report?
Two questions, and both have to be yes. First, was the report produced by a laboratory that does not work for the brand. Second, does it carry the lot or batch code printed on the bottle in your hand. A report keyed to a product code or a barcode describes the product in general, not your bottle, and in house testing is the company checking its own work. GC/MS tested printed on a label answers neither question.
What can I use if synthetic fragrance makes me ill?
Unscented is the only version of a product whose scent chemistry is not a question, so where an unscented option exists it is the answer. Beyond that, two disclosure tools exist. California's Cosmetic Fragrance and Flavor Ingredient Right to Know Act makes companies report fragrance and flavour ingredients appearing on 23 hazard lists to a public, searchable state database. And since 31 July 2026 any cosmetic newly placed on the EU market must name 81 fragrance allergens on the label rather than 26, so an EU labelled pack of the same product tells you considerably more.
Do essential oils damage plastic?
Yes, and citrus oils are the clearest case. D-limonene makes up roughly 60 to 97 percent of cold pressed citrus oils, and it is the same compound sold industrially as a solvent for dissolving polystyrene. Neat essential oil will craze and soften polystyrene, swell and distort PET, and stress crack thinner polyethylene. Keep oils in glass, never decant them into a plastic bottle, and do not drip them into a plastic diffuser or humidifier tank.
Why do wintergreen essential oils keep getting recalled?
Wintergreen and birch oils are mostly methyl salicylate, and United States law requires any product holding more than a small share of it to be sold in child resistant packaging under the Poison Prevention Packaging Act. A steady run of recalls since 2019 has been for exactly this: the oil was fine, the cap was not. One 2021 recall covered about 1.3 million bottles. If you have small children and keep wintergreen or birch oil, check that the cap is genuinely child resistant.
If you remember one thing, make it the two questions. Not "is it pure", which nothing on the label can settle, but "who ran the report" and "does it carry my lot code". Those two answers tell you whether anyone has actually measured the bottle you are holding.
Related Articles
Want a verdict on a brand that is not here? Product Check will look it up, and research it if nobody has.