Beginner perfumer · 68
How do you tell a material has been retired — the database has two fields for it, and they overlap on 18 records
· 14 min read
When you look up an obscure material, what you want to know is whether it can still be used. This database answers in two places: the category field can read "information only not used for fragrances or flavors", and the cosmetic_uses field can read "not used anymore". The first covers 6,457 records (15.3% of the database); the second covers 80. Both appear together on 18. Worse, of those 80 "not used anymore" records, 62 still carry a flavour or fragrance category. Dihydrocoumarin has 55 suppliers, allyl isothiocyanate 53, costus root oil 21 — all of them wearing both contradictory labels at once.
About 5 minutes.
When you look up a material you have never heard of, the first thing you want is usually not what it smells like. It is whether you can still use it.
This database has two fields that answer that. They often disagree.
Two retirement flags
The first is in category, with the value "information only not used for fragrances or flavors".
Of the 42,225 records, 6,457 carry it — 15.3%, one in six.
The second is in cosmetic_uses, with the value "not used anymore".
That one covers 80 records.
Both together: 18.
category = information only 6,457 records
cosmetic_uses = not used anymore 80 records
both 18 records
One is a sixth of the database, the other is two in ten thousand. They are plainly describing different things.
The first field is not about retirement
A large share of those 6,457 records are not fragrance materials at all. Pull the ones that still have ten or more suppliers and you get 18:
| Material | Suppliers | What it is |
|---|---|---|
| Iodine | 21 | An element |
| Musk xylol | 20 | A nitro musk |
| Geranyl nitrile | 20 | A citrus material |
| Musk ambrette | 18 | A nitro musk |
| Sodium cyclamate | 15 | A sweetener |
| Lactose | 13 | A sugar |
| Bromstyrol | 13 | A green material |
| Potassium ferricyanide | 12 | A reagent |
| Barium carbonate | 10 | A reagent |
What this field holds is "the database carries this record, but it is outside the fragrance and flavour scope". Iodine, lactose and barium carbonate were never fragrance materials. They are not retired; they never enlisted.
The genuinely retired ones are mixed in among them — musk xylol, musk ambrette, geranyl nitrile, bromstyrol.
So this field cannot be used as a retirement flag. When it lights up, ask first: was this ever a fragrance material?
The second field is about retirement, but it fights with category
Of those 80 "not used anymore" records:
- 72 still have suppliers
- 62 still carry a flavour or fragrance category
That is two fields on the same record contradicting each other: one says it is not used anymore, the other classifies it as a flavour or fragrance agent.
The ones with five or more suppliers:
| Material | Suppliers | category still reads |
|---|---|---|
| Dihydrocoumarin | 55 | flavor and fragrance agents |
| Allyl isothiocyanate | 53 | flavoring agents |
| Massoia lactone | 32 | flavoring agents |
| (E)-2-heptenal | 31 | flavoring agents |
| Furfuryl alcohol | 31 | flavoring agents |
| Phenol | 28 | food contact materials |
| Dimethyl sulfoxide | 24 | multipurpose additives |
| Costus root oil | 21 | flavoring agents |
| Musk ambrette | 18 | information only |
Dihydrocoumarin has 55 suppliers and is flagged "not used anymore".
That is not necessarily an error. Both can be true at once: some jurisdiction or some use was closed off, and it still sells elsewhere. Costus root oil is exactly that — restricted on sensitisation grounds, but still a purchasable essential oil.
But it means this field answers "is it on some list", not "can I use it" — and it does not say which list.
A worked case of the two flags disagreeing
Take the two nitro musks this site has written about:
| Musk xylol | Musk ambrette | |
|---|---|---|
| CAS | 81-15-2 | 83-66-9 |
| category | information only | information only |
| cosmetic_uses | fragrance; perfuming agents | not used anymore |
| Suppliers | 20 | 18 |
| Odour type | musk | musk |
Two equally retired materials, and one has cosmetic_uses reading "not used anymore" while the other reads "fragrance; perfuming agents".
Musk xylol is also one of only two records in all 42,225 that are simultaneously categorised "information only" and have "fragrance" in cosmetic_uses. The other is sodium cyclamate, an artificial sweetener.
(The full record is in Material guide: musk xylol.)
So what should you check instead
Neither field is a regulatory source. They are this database's housekeeping columns, and the housekeeping is inconsistent. To actually establish whether a material can be used, in order:
- The IFRA standards. This is the industry's own framework, and it says whether something is prohibited or restricted and for which product categories. How a limit is calculated is taken apart in Citral.
- Your own jurisdiction's cosmetic regulation. EU CosIng and Annex II, Japan's cosmetic standards, and so on. The same material can get different answers in different places.
- The supplier's own SDS and specification sheet. If they sell it to you, restrictions on use are theirs to state.
- These two database fields — as a prompt, not a conclusion. If either lights up, go and check the three above.
Supplier count is no guide at all. Dihydrocoumarin at 55 and musk xylol at 20 are both still being sold. Buyable and usable are different things, which I also ran into in Deciding what to buy from how many formulas use it.
What this doesn't establish
- This is an audit of this database's fields, not a regulatory check. Every number above is about how fields are filled in, not about what any law says.
- I did not trace which list stopped each of the 80 "not used anymore" records. Some will be IFRA, some food regulation, some possibly just discontinued production.
- "62 still carry a fragrance category" is not necessarily an error. A material can be dropped from cosmetics while remaining usable as a flavour; the two fields describe different uses.
- Of the 6,457 records, I only classified the 18 with ten or more suppliers. The other 6,439 I did not sort into "retired material" versus "never a material".
- The fields may be out of date. I do not know when these markings were last revised.
- No data does not mean safe. The empty-field problem is counted in Four kinds of nothing found.