"I want an environmentally friendly, non-hormone-disrupting musk" — the honest answer is worse than it sounds

Someone posted a clear brief on r/DIYfragrance. They wanted two or three more musks, and listed what they had tried:

"Galaxolide: Love. But I'm hoping to find an environmentally friendly, non-hormone disrupting alternative. Wouldn't repurchase."

"Ethylene Brassylate: Smelled oily, and underwhelming to me. Didn't love, and wouldn't repurchase."

"Habanolide: Very interesting hot, metallic laundry. 7/10. Would repurchase."

Neither of those two reasons is idle. Both can be checked.

And checking them does not produce "buy this one instead".

The short version

• The database is silent about most musks. There are 36 musk entries with ten or more suppliers, and 23 of them (64%) have no GHS field at all.
• Of the 13 that do have one, 8 carry an aquatic hazard — and all 8 are polycyclic or nitro musks, five of them Galaxolide at different dilutions.
• "Nothing recorded" is not "nothing found". Those 23 blanks are mostly unfilled rather than tested-and-cleared.
• A 2023 global review stops in the same place. It states plainly that data on macrocyclic and alicyclic musks are limited for both occurrence and PBT properties.
• The same review offers a counterintuitive number: risk quotients for HHCB, AHTN, musk xylene and musk ketone are below 0.1 in most waters and sediments, a low risk, with RQ>1 only at some sites such as those near sewage treatment plants.
• The hormone claim rests on research more complicated than the slogan. A 2002 study found HHCB and AHTN act as selective estrogen receptor modulators, with direction depending on cell line and receptor subtype, and weak estrogenic effects only at 10 μM. The authors' own conclusion: caution should be exerted when interpreting such effects from in vitro systems.
• So the honest answer: the alternatives are not demonstrably better, they are less studied.

About 9 minutes.

What the database records about musks

Start with the shape of the data. Entries whose odour type contains musk and which have ten or more suppliers — 36 of them:

	Entries
All	36
No GHS field at all	23 (64%)
With a GHS field	13
Of those, carrying an aquatic hazard	8

Those 8:

Suppliers	Name	Aquatic
38	musk indanone	H411 toxic
34	musk gx 100% (Galaxolide / HHCB)	H400 / H410 very toxic
32	musk tetralin	H400 / H410
30	musk gx 50% in IPM	H400 / H410
27	musk gx 50% in DEP	H400 / H410
22	musk gx 50% in BB	H400 / H410
20	musk xylol	H410
14	musk gx 50% in DPG	H400 / H410

Five of the eight are Galaxolide at different dilutions. And all eight sit on the polycyclic or nitro side.

That person's instinct was right. Galaxolide does carry the heaviest aquatic labelling in the whole musk group.

The important column is the other one. The five entries with a GHS field but no aquatic statement look like this:

Suppliers	Name	GHS
52	ethylene brassylate	H315 skin irritation
52	ω-pentadecalactone	H302 harmful if swallowed
43	isoambrettolide	H315, H319

On the macrocyclic side, the labels are about irritation rather than water.

Do not read that as proof. Many of those 23 blank records are macrocyclic musks, and a blank means "no data here" rather than "tested and fine".

The 2023 review stops in the same place

Wang and colleagues reviewed the global distribution and ecological risk of synthetic musks in Environmental Pollution in 2023.

The picture has three layers:

First, which ones get detected.

Galaxolide (HHCB), Tonalide (AHTN), musk xylene (MX) and musk ketone (MK) are generally the most frequently detected synthetic musks in different samples, with HHCB and AHTN being predominant.

Concentrations of HHCB and AHTN run higher in western countries than in Asian ones, which the authors read as a difference in how much gets used.

Second, how high the risk is. This layer is milder than most people expect:

The risk quotients of HHCB, AHTN, MX and MK in most waters and sediments are below 0.1, reflecting a low risk to aqueous and sediment-dwelling species. At some sites, for example close to sewage treatment plants, high risks (RQ>1) are characterised.

"Most frequently detected" and "highest risk" are different things. They get detected because volumes are large and the analytical methods are mature.

Third, and this is the point of the article:

Currently, limited data are available for macrocyclic musks and alicyclic musks in terms of either occurrence or PBT properties.

That sentence and the database's 23 blanks are the same fact.

Two entirely independent sources — a commercial materials database and a 2023 environmental review — go silent at the same point, for the same reason: nobody has measured.

So "switching to macrocyclic musks is greener" is not a demonstrated statement at present. It is a reasonable untested inference. There is a structural argument for it (macrocyclic esters should biodegrade more readily than polycyclic aromatics), and an argument is not data.

What the hormone research actually measured

This is the most widely repeated and most heavily compressed claim on the internet. The original work is a good deal more complicated.

Schreurs and colleagues used in vitro reporter gene assays in Toxicology and Applied Pharmacology in 2002 to test AHTN and HHCB against estrogen receptors.

What they found:

• Both act as selective estrogen receptor modulators (SERMs), inducing both estrogenic and antiestrogenic activity, with the direction depending on the cell line and the ER subtype targeted.
• Weak estrogenic effects were observed only at relatively high concentrations (10 μM).
• Antiestrogenic effects appeared in various cell lines from 0.1 μM.
• Compared with the well-known SERM 4-hydroxytamoxifen, AHTN and HHCB have a much lower potency in suppressing estradiol-induced transactivation.

And their own conclusion:

The cell type dependency of ER transactivation shows that caution should be exerted when interpreting effects of estrogenic compounds using in vitro systems.

All four of those hold at once: the interaction is real, its direction changes with the system, the potency is far below the reference compound, and the authors themselves warn against over-reading.

"Hormone disrupting" compresses those four into one phrase. We've written about reading fragrance safety studies, and this is the same pattern: in vitro receptor activity is a signal, not a dose.

None of which says there is nothing there. That paper is from 2002 and twenty-odd years of work followed it. What it says is that if this is the reason you are switching materials, the alternatives have no matching body of research saying they do not do the same thing — because that research does not exist yet.

The corpus: how much each of these is actually used

Across the 954 public formulas:

Material	Formulas	Median dose
musk ketone (nitro)	84 (8.8%)	4.00%
ethylene brassylate (macrocyclic)	59 (6.2%)	8.33%
Habanolide (macrocyclic)	35 (3.7%)	5.00%
Ambrettolide (macrocyclic)	30 (3.1%)	1.45%
Helvetolide (alicyclic)	15 (1.6%)	2.78%
Exaltolide (macrocyclic)	13 (1.4%)	3.00%
Galaxolide (polycyclic)	7 (0.7%)	9.00%

Two things to notice.

First, musk ketone appears in 84 formulas, twelve times as often as Galaxolide. That is not today's market, that is this corpus's vintage — 74% of it comes from patents and published sources, where nitro musks were still common. We've written about the corpus's selection bias, and this is its most visible instance. The table cannot be read as current practice.

Second, the one they dismissed is the most used macrocyclic in the set.

Their verdict on ethylene brassylate was oily and underwhelming. In this corpus it is the most frequently appearing macrocyclic musk (59 formulas), at a median dose of 8.33%.

"Underwhelming" may be a dosing problem. Its substantivity is 208 hours at 100%, short for a musk, and a median of 8.33% suggests formulators treat it as a volume material. If they evaluated it at 1-2%, they did not smell what it does in a formula.

(To be clear: I have not smelled it. That is a possible explanation derived from numbers, not a diagnosis.)

So what should they actually do

Collecting the above into something actionable.

1. If the reason is aquatic, the data supports avoiding polycyclic and nitro musks. All eight aquatic labels fall on that side, and all four of the most-detected musks in Wang's review are there too. This step has ground under it.

2. Do not call the alternative "safe" — call it "untested". That is the honest position, and it changes how you use the material: something untested should not get dosed higher just because you switched to it.

3. If the reason is endocrine, that reason cannot help you choose. There is no matching research on the alternatives to compare against. You are moving from the side with data to the side without it.

4. What actually changes is the formula, not the environment. Macrocyclic and polycyclic musks are not one-for-one in odour or in dose:

	Galaxolide (polycyclic)	ethylene brassylate (macrocyclic)
Substantivity	400 hours at 100%	208 hours at 100%
Corpus median dose	9.00%	8.33%
Suppliers	34	52

Nearly half the substantivity. After switching, the drydown gets shorter and something else has to carry it. That is the real cost of this change, and it has nothing to do with the environment.

5. They already did one thing right: they listed every material they had tried, with a verdict. That list beats anybody's recommendation, because it is data their own nose produced. We've written about keeping evaluation notes.

What this doesn't establish

I have smelled none of these musks and made no comparisons. The whole piece sets database fields, the formula corpus and two papers against each other.

"A blank GHS field means nobody filled it in" is speculation. The database does not explain why a field is empty, and those 23 blanks could have other causes. The one thing I can be sure of: a blank is not a safety conclusion.

The eight aquatic labels falling on the polycyclic and nitro side is this database's distribution, not a complete regulatory list. If a macrocyclic musk carries an aquatic classification elsewhere, I cannot see it from here.

Wang's paper is a review, not primary measurement. The sub-0.1 risk quotients were compiled from existing studies, and an RQ depends on the effect concentration and exposure estimate adopted. Different assumptions give different quotients.

I found no direct comparative study for "macrocyclics biodegrade better". It is a structurally coherent inference, and Wang's review says explicitly that this area is data-poor. I have written it as speculation because that is what it is.

Schreurs is a 2002 in vitro study. Twenty-odd years of literature followed, and I read only that one, so this piece presents that result's content and limits rather than a summary of the field's current state.

10 μM is a high concentration. I cannot convert it into "how much you get from wearing perfume", because that needs an exposure model and there is none in that paper.

The corpus's vintage bias is severe here. Musk ketone at 84 formulas against Galaxolide at 7 reflects the age of the literature rather than the market. That table can show relative dosing and not who is popular now.

The two substantivity figures (400 against 208 hours) come from two database records, both marked 100%, with no guarantee of one experiment or one method.

References

T. Wang, H. Zou, D. Li, J. Gao, Q. Bu, Z. Wang, Global distribution and ecological risk assessment of synthetic musks in the environment, Environmental Pollution, 331(Pt 2), 121893 (2023). PMID 37245793. doi:10.1016/j.envpol.2023.121893

R. H. Schreurs, M. E. Quaedackers, W. Seinen, B. van der Burg, Transcriptional activation of estrogen receptor ERalpha and ERbeta by polycyclic musks is cell type dependent, Toxicology and Applied Pharmacology, 183(1), 1-9 (2002). PMID 12217637. doi:10.1006/taap.2002.9458

Related: synthetic musks are not one family, how to read fragrance safety studies, keeping evaluation notes, how big is the palette.