Material guide: allyl isothiocyanate — the notes field says 'used in the manufacture of flavors, war gases,' and it does not exist in the intact plant

About a 5 minute read.

The short version

• Name: allyl isothiocyanate, CAS 57-06-7, FEMA 2034
• Odour type: sulfurous; the odour field holds two words — pungent, mustard
• Evaluated at 0.01% — the lowest tier among the database's thinly-described records (vanilla extract sits at 100%)
• Suppliers: 53; boiling point 150–152 °C; shelf life 18 months
• Dermal LD50 88 mg/kg (rabbit), subcutaneous rat 92, subcutaneous mouse 80
• cosmetic_uses: "not used anymore"
• It does not exist in the intact plant

The first sentence of the notes field

used in the manufacture of flavors, war gases; medical use as a counterirritant. Chief constituent of natural mustard oil. Also found in cooked cabbage, horseradish, etc.

"War gases," in the notes field of a flavour ingredient database.

Not an exaggeration — isothiocyanates are lachrymatory and vesicant, and they were genuinely investigated around both world wars. And in the same sentence it is a FEMA GRAS flavouring. One molecule, one line of one document.

(cosmetic_uses reads "not used anymore" — a field I measured in two retirement flags. This one was not banned; the cosmetic side simply stopped using it.)

It does not exist in the intact plant

The occurrence field lists garlic mustard, black mustard, broccoli leaf, brussels sprout seed, cabbage leaf, horseradish root, mustard greens seed, rocket, wasabi.

But before you cut them, none of those plants contains this molecule.

Dosz and colleagues, in the Journal of Agricultural and Food Chemistry in 2014, were working on exactly this mechanism:

Isothiocyanates, generated from the hydrolysis of glucosinolates in plants of the Brassicaceae family … Hydrolysis requires the plant enzyme myrosinase.

The glucosinolate (sinigrin, in mustard) and myrosinase live in separate compartments. While the cell is intact they never meet. The moment you cut, grate or chew, the compartments break, the enzyme finds its substrate, and allyl isothiocyanate gets made.

This is a plant defence system: keep the weapon in two halves and assemble it when bitten.

Their method is interesting in itself — they used AITC formation to estimate myrosinase activity, and did it on unextracted fresh tissue, because the conventional approach of isolating crude protein can underestimate activity in a whole food.

(This is the same fact in a different key as the line from birch tar: the molecule you smell was not in the raw material; a process made it. There it was heat. Here it is an enzyme.)

It shares neurons with chilli and cannabis

The question Jordt and colleagues took up in Nature in 2004 was direct: how does mustard heat work?

Wasabi, horseradish and mustard owe their pungency to isothiocyanate compounds. Topical application of mustard oil (allyl isothiocyanate) to the skin activates underlying sensory nerve endings, thereby producing pain, inflammation and robust hypersensitivity to thermal and mechanical stimuli. Despite their widespread use in both the kitchen and the laboratory, the molecular mechanism through which isothiocyanates mediate their effects remains unknown.

And the result:

we show that mustard oil depolarizes a subpopulation of primary sensory neurons that are also activated by capsaicin, the pungent ingredient in chilli peppers, and by Δ9-tetrahydrocannabinol (THC), the psychoactive component of marijuana. Both allyl isothiocyanate and THC mediate their excitatory effects by activating ANKTM1, a member of the TRP ion channel family recently implicated in the detection of noxious cold.

Mustard, chilli and cannabis — three entirely different plants hitting the same population of neurons. And mustard and THC go further, through the same channel.

That channel was thought to handle painful cold. So the sensation of mustard shooting up your nose runs on the same wiring as hands hurting in the cold.

(Capsaicin works through a different channel, TRPV1, which handles painful heat — so the intuition that chilli heat is hot and mustard heat is cold has a structural reason behind it.)

Why it is evaluated at 0.01%

The odor_descriptions field:

at 0.01% in propylene glycol. strong pungent mustard

I measured this last time: across the 56 records with one or two descriptors and twenty-plus suppliers, evaluation concentrations span 0.01% to 100%, a ten-thousand-fold range. And only three sit at 0.01% — this material and two garlic oils.

It is the bottom of that scale.

The toxicity field explains why:

	Value
Oral (rat)	LD50 339 mg/kg (Jenner 1964) / 490 (Vernot 1977)
Dermal (rabbit)	LD50 88 mg/kg
Subcutaneous (rat/mouse)	92 / 80 mg/kg
Intraperitoneal (mouse)	LDLo 4 mg/kg

88 mg/kg dermal. Most aroma materials' rabbit dermal LD50 reads "> 5000 mg/kg." When I wrote up isovaleric acid I called its converted 287 mg/kg low — this is more than three times lower again.

This is not a perfume material. Its category field is flavoring agents, and its cosmetic use is "not used anymore."

On the taste side

The flavor field: mustard, horseradish, wasabi.

taste_descriptions holds five entries, one of them with an application note:

pungent, sharp, horseradish. Useful in: savory vegetable, savory spices, savory allium, savory others.

That is its real job. Japanese mustard, horseradish sauce, prepared mustard — the sensation that goes up your nose is this.

How to use it (if you really are going to)

• Treat it as a flavour material, not a fragrance one. FEMA 2034, and the cosmetic side is done with it.
• Evaluate at 0.01%. The lowest tier in the database.
• Dermal LD50 88 mg/kg. Keep it off skin. It is a counterirritant; that effect is the design.
• What it does in your nose is trigeminal pain, not olfaction. That rush is not odour intensity, it is pain.
• 18 months, cool and dry.

What this doesn't establish

• I have not smelled it (and have no intention of doing so above 0.01%). This is fields, whole-database statistics, and two papers.
• Neither paper is fragrance research — Jordt is sensory neurobiology, Dosz is enzyme activity in produce. I cite their account of this molecule's mechanism and origin.
• "War gases" is the notes field's claim; I checked no historical source.
• The line about capsaicin and TRPV1 is prior knowledge, not drawn from either paper.
• I did not go to the source for the intraperitoneal LDLo of 4 mg/kg (Toxicology and Applied Pharmacology, 1972).
• I met the source "Jenner 1964" once before as a record with a suspect unit in the acetophenone article; this record reads mg/kg, so the unit itself is fine here.
• I checked no IFRA or other regulatory text.