Material guide: allyl amyl glycolate — change the solvent and its no-effect level drops to a third

CAS 67634-00-8. Forty-six suppliers, 58 of 954 formulas, median dose 0.67%. A 2006 cumulative irritation study in 129 human subjects did something rarely done: it tested one material in three vehicles. In diethyl phthalate its no-observed-effect level came out at 0.33%; in an ethanol-rich 1:3 vehicle it fell to 0.12% — same material, same subjects, nearly threefold. And perfume's vehicle is ethanol. Even finding the paper is a trap: this material's RIFM assessment is filed under allyl (3-methylbutoxy)acetate.

This series has written up plenty of "what is the limit for this material", and almost every time that limit was a number.

This one demonstrates that the number depends on what you dissolved it in.

A 2006 study in 129 human subjects tested one material in three vehicles, and the no-observed-effect level fell from 0.33% to 0.12%.

The vehicle it fell in was the ethanol-rich one.

The short version

CAS 67634-00-8, 46 suppliers, 58 of 954 formulas, median dose 0.67%.
Odour: fruity, green, galbanum, pineapple, fusel. Strength medium, recommend smelling at 10% or less. Substantivity 84 hours at 100%.
It is an allyl ester. Allyl esters often contain a certain percentage of free allyl alcohol, which is known to have a potential for delayed skin irritation.
Politano and colleagues ran a cumulative irritation test in 129 subjects in 2006, putting five allyl esters including this one into three vehicles: neat DEP, 3:1 DEP:ethanol, and 1:3 DEP:ethanol.
This material's no-observed-effect level: 0.33% in DEP, 0.12% in 1:3 DEP:ethanol.
The paper's summary: higher levels of ethanol in the vehicle showed a trend toward lower concentration thresholds to induce irritation.
This material's RIFM safety assessment is filed as allyl (3-methylbutoxy)acetate — a name that matches nothing, with only the CAS in common.

About 8 minutes.

The vehicle changes the answer

The study first, because it is this piece's spine.

Politano and colleagues published in the International Journal of Toxicology in 2006 with a clear question: allyl esters are frequently used in the fragrance industry and often contain a certain percentage of free allyl alcohol, which is known to have a potential for delayed skin irritation. Finished products also contain different solvent systems, or vehicles, chosen for the intended application. Do those vehicles affect the skin irritation potential?

The method:

Five allyl esters: allyl amyl glycolate, allyl caproate, allyl (cyclohexyloxy)acetate, allyl cyclohexylpropionate, allyl phenoxyacetate
Three vehicles: diethyl phthalate (DEP), 3:1 DEP:ethanol, 1:3 DEP:ethanol
Four concentrations: 0.1%, 0.5%, 1.0%, 2.0% (w/w)
129 human subjects, modified cumulative irritation test
0.3 ml applied to the back under occlusion for 24 hours using Hill Top chambers
Irritation assessed at 1, 2, 4 and 5 days after application

The results:

Cumulative irritation scores varied considerably among test materials. There were no delayed irritation observations. The highest irritation scores were observed at the 2.0% concentration for all test materials.

And the vehicle comparison is the part that matters:

The irritation scores for allyl amyl glycolate, allyl (cyclohexyloxy)acetate, and allyl phenoxyacetate were highest in 1:3 diethyl phthalate:ethanol, thus the resulting calculated no-observed-effect levels, 0.12%, 0.03%, and 0%, respectively, were much lower for this vehicle compared to the diethyl phthalate vehicle, 0.33%, 0.26%, 0.25%.

These data showed a trend for lower concentration thresholds to induce irritation when higher levels of ethanol were used in the vehicle.

One material, the same 129 people, a different solvent, and the threshold moves by a factor of 2.75.

Allyl phenoxyacetate is more dramatic still: from 0.25% to 0%.

Why this matters particularly when you make your own

Because perfume's vehicle is ethanol.

Of the three tested, 1:3 DEP:ethanol has the most ethanol in it, and an alcoholic perfume has more than that.

This does not mean your formula will cause trouble at those numbers. That study used 24-hour occluded patches — not the same event as spraying your wrist. Occlusion raises absorption substantially, and a perfume is applied openly and evaporates.

What it establishes is a general idea: a material's safe use level is not a property of the material but of material-plus-vehicle.

That is easy to lose when reading a datasheet. An SDS or a limits table hands you a percentage, and behind that percentage is an assumed product form. IFRA's limits are organised by product category, which is the institutional version of exactly this point.

In practice:

Diluting in DPG to evaluate, then putting it into an alcoholic perfume, is a change of vehicle.
Copying a percentage out of somebody else's formula may also be a change of vehicle.
This material's corpus median is 0.67% of a concentrate. Diluted to a 20% finished product that is 0.13% — almost exactly the 0.12% calculated in the ethanol-rich vehicle.

That last comparison needs its limits stated. 0.67% is a share of concentrate, 0.13% is my conversion using a common dilution rate, and 0.12% is a no-effect level from a 24-hour occluded patch. All three sit under different conditions, and the near-match is a coincidence of coincidences. I write it out because it makes you stop and think, not because it demonstrates anything.

What the material is

Property	Value
CAS	67634-00-8
Formula	C₁₀H₁₈O₃, MW 186.25
Appearance	colourless clear liquid (est)
Boiling point	201-203 °C
Flash point	> 93.3 °C
logP	2.30 (est)
Water solubility	510.7 mg/L at 25 °C (est)
Shelf life	24 months or longer
Substantivity	84 hours at 100%
Strength	medium, recommend smelling at 10% or less
Suppliers	46
Occurrence	not found in nature
Odour	fruity, green, galbanum, pineapple, fusel
GHS	H315, H319, H335

"Galbanum plus pineapple" is an unusual pairing.

Galbanum is the sharpest thing in the green family and pineapple is tropical fruit. This material gives you both at once, which is its job in a formula: pulling a fruit toward green, or pulling a green toward something friendlier.

Substantivity of 84 hours is short for this series — far short of phenylacetaldehyde's 400 and far longer than dimethyl sulfide's 4. It works in the middle.

The 0.67% median dose is worth noting. We built the ladder from strength label to dose: materials labelled medium sit at a median of 2.00%. This one runs at a third of that, and its strength label carries the extra "recommend smelling at 10% or less" — the database's way of marking something as needing more care than an ordinary medium.

The name does not match

This material demonstrates a very practical problem: you cannot find its safety assessment by name.

Its RIFM ingredient safety assessment is titled:

RIFM fragrance ingredient safety assessment, allyl (3-methylbutoxy)acetate, CAS Registry Number 67634-00-8

Not one word of "allyl amyl glycolate" appears.

Trade or common name: allyl amyl glycolate (sometimes AAG)
IUPAC-style registered name: allyl (3-methylbutoxy)acetate
CAS: 67634-00-8 ← the only thing that matches

The same problem as the trade names in the jasmine piece and the RIFM mix-up in the amyl cinnamal piece: check the CAS, not the name.

And that RIFM assessment has no abstract on PubMed. I can confirm it exists and matches the CAS, and I cite none of its conclusions. It is in the references so you know where to look, not so I can relay what I have not read.

What this doesn't establish

I have not smelled it. The whole piece sets database fields, the formula corpus and one paper against each other.

Politano's study used 24-hour occluded patches, the standard design for an irritation test and very far from how a perfume is used. Occlusion raises absorption, 24 hours far exceeds ordinary contact, and perfume evaporates. Those no-effect levels cannot be read as safe limits in a perfume.

That study is RIFM's own work (the authors are at the Research Institute for Fragrance Materials). As with the environmental study in the companion piece, I state the affiliation so readers know, not to question the design — the three-vehicle comparison is solidly built.

The 0.67% and 0.12% comparison sets three differently-conditioned numbers side by side. Flagged in the body and repeated here: that is not a safety assessment.

"Free allyl alcohol" is the paper's statement. Free allyl alcohol content will differ by supplier and by batch, and the database has no field for it. That means this material's irritation potential may vary between batches, and I cannot quantify that.

For the 2023 RIFM assessment I confirmed only its existence and CAS. There is no abstract on PubMed, and I do not relay what I have not read.

The GHS classification is what the database records (H315, H319, H335), not a complete regulatory determination. Work from the SDS for your own batch and the current IFRA standard.

References

V. T. Politano, D. A. Isola, J. Lalko, A. M. Api, The effects of vehicles on the human dermal irritation potentials of allyl esters, International Journal of Toxicology, 25(3), 183-193 (2006). PMID 16717034. doi:10.1080/10915810600683275

A. M. Api, D. Belsito, D. Botelho, et al., RIFM fragrance ingredient safety assessment, allyl (3-methylbutoxy)acetate, CAS Registry Number 67634-00-8, Food and Chemical Toxicology, 182 Suppl 1, 114226 (2023). PMID 38012994. doi:10.1016/j.fct.2023.114226