A fixative does not hold anything down: why musks last, and why what lasts is not your formula

"Add a little musk or benzoin and it will last" mixes two effects. The musk itself may remain for a long time; by then, the faster materials from the original formula have still left. Fixation reshapes an evaporation curve. It cannot pause the top notes.

Evaporation is an equilibrium problem, not a gripping problem

The word "fixative" invites you to picture glue: one molecule holding the others so they cannot escape. The picture is intuitive. The mechanism works differently.

In 2021 Rodrigues, Nogueira and Faria reviewed two decades of work from the Laboratory of Separation and Reaction Engineering at the University of Porto in Molecules. Their evaporation and one-dimensional diffusion model rests on two things: vapour-liquid equilibrium and Fick's law of diffusion. The crucial step, in their words, is predicting vapour composition "using group-contribution methods (UNIFAC, UNIFAC-D, ASOG, A-UNIFAC, COSMO-SAC) for calculating the activity coefficients", with a modified Raoult's law connecting liquid composition to gas-phase partial pressures.

How fast each component evaporates at any moment depends on its own vapour pressure, its mole fraction in the mixture, and its interactions with everything else (the activity coefficient). Search the model for a "grip" term and you will come up empty. Equilibrium does all the work.

Two further conclusions in that review bear directly on fixation:

1. Odour value decides what you smell. Their Perfumery Ternary Diagram is plotted using odour values, with Perfumery Binary Lines marking where two adjacent components have equal odour value, and this determines the dominant odour of each composition. The conclusion from the article on odour thresholds is formalised here: dominance belongs to whatever has the greatest partial pressure relative to its threshold, however little of it you poured in.

2. The matrix changes the sequence of dominance. They report that the matrix effect substantially influences release (base notes in dipropylene glycol show high retention) and that different matrices materially change which odour dominates over time. The previous article on solvents arrives at the same place from another direction.

So a fixative really does two things: it dilutes (lowering the mole fraction of everything else) and it shifts activity coefficients. Both effects are real. Both are also modest, far short of the longevity beginners are hoping for.

Then why does adding a musk genuinely feel like it lasts longer?

Musk really is the most substantive family

Start with the data. Taking materials with supply information, a primary odour family, and a recorded substantivity:

Odour family	With a substantivity value	Of those, at 400 hours	Median of the rest
Musk	36	53%	336 hours
Animal	23	52%	168 hours
Spicy	84	13%	116 hours
Woody	120	39%	112 hours
Balsamic	94	45%	100 hours
Floral	316	12%	68 hours
Citrus	109	4%	28 hours
Minty	38	0%	20 hours

(Overall median excluding 400 hours: 53 hours, n=1,295.)

How to read this table. 400 hours is the single most common value in this field — a conventional top of scale rather than a measurement, as the article on evaporation curves established. So we report the share sitting at 400 hours and the median of everything else separately: the first is itself informative (it means "does not meaningfully evaporate on the timescale of the test") and only the second is a comparable measurement.

Reading the median column alone would understate musk, animal, woody and balsamic, because 39–53% of their values sit at the ceiling and were excluded. Both columns have to be read together.

Coverage matters too. Substantivity is recorded for only part of each family: 36 of 65 musks (55%), 120 of 354 woody materials (34%), and just 84 of 375 spicy ones (22%). With n=36, musk here is directional evidence, not a precise estimate.

Even with every caveat applied, the direction is unmistakable: musk is the most substantive family, by a wide margin. 336 hours against 53 overall.

So the first half of the advice (musks themselves last) is correct.

But lasting and preserving the composition are different things

The problem is the second half.

If you can still smell something at hour eight, what is that something?

By the logic above: the dominant odour at any moment is the component with the highest odour value. Your bergamot left long ago, your jasmine is nearly gone, and what remains is the musk. At hour eight the musk is what reaches anyone's nose; the perfume you composed has already left the stage.

This is longevity by substitution. The thing that made it to hour eight is your last-departing ingredient. The composition it once belonged to ended hours earlier.

There is a second trap stacked on top, from the article on olfactory fatigue: adaptation is stimulus-specific. Wearing your own perfume for eight hours adapts you to it, so you conclude it has faded while someone next to you still finds it clear. "My perfume doesn't last" is sometimes a fatigue problem wearing an evaporation costume. Judge longevity by asking someone else, or on a strip. I write the dip time on the tail of every strip and revisit it through the day; the strip has corrected my wrist more than once.

A musk base needs more than one nose

OR5AN1 genetics alters the perceived strength of some macrocyclic musks. That affects fixation trials, but musk anosmia is a separate lesson; see Synthetic musks are not one family.

A relationship we computed and could not use

A negative result worth recording.

We noticed an attractive pattern: materials with an odour strength of low have a median substantivity of 322 hours, against 88 hours for medium. It looks like "weaker things last longer", and it tells well.

It is almost entirely an artifact of that 400-hour ceiling. In the low group 43% of values sit at 400 hours; in the none group it is 71%, and that group is largely solvents such as DPG and IPM.

Excluding 400 hours:

Odour strength	n	Median substantivity
low	30	150 hours
medium	1,014	48 hours
high	224	72 hours

The pattern breaks immediately: high outlasts medium, which kills any monotonic story, and low is down to 30 samples.

So "weaker lasts longer" does not hold, and we did not use it.

So how should you actually approach fixation?

1. There is no odourless fixative. The table already answers this: substantive materials cluster in musk, animal, woody, balsamic and spicy, all families with strong characters. Choosing a fixative is choosing a smell. The only odourless non-volatiles are solvents like DPG, and their contribution is dilution (previous article).

2. Decide what you want left behind. Hour eight belongs to your most substantive component, so design backwards. Pick the smell you want to remain at the end, and choose your fixative from there rather than reaching for a musk on reflex.

3. Cover every stretch of the curve. Evaporation is a continuous curve, as the evaporation-curve article argued, and fixation means leaving no gap in it. Give the heart a few descriptors in common with the base so the handover holds.

4. Add low-volatility materials early and macerate everything together. Same conclusion as the evaporation-curve article.

5. Verify on a strip and with other people's noses. Adaptation will make you systematically underestimate your own formula's longevity; your wrist is the least reliable instrument you own.

6. Write down the 24-hour point. The note format from the record-keeping article has that field, and this article is its purpose. The three to five descriptors you record at 24 hours are what your formula actually leaves behind. If they have nothing to do with what you set out to make, your fixation is aimed wrong, and only the notes will show you.

---

Fixation is not extension. It is deciding who is still on stage at the end. Adding a musk genuinely makes something last longer, but that something is the musk. Decide what you want to last first; how to make it last is the easier half of the problem.

References

A. E. Rodrigues, I. Nogueira & R. P. V. Faria, Perfume and flavor engineering: a chemical engineering perspective, Molecules, 26(11), 3095 (2021). PMID 34067262

N. Sato-Akuhara et al., Genetic variation in the human olfactory receptor OR5AN1 associates with the perception of musks, Chemical Senses, 48, bjac037 (2023). PMID 36625229

---

Next up, individual differences: if everyone's receptors differ, is there any consensus about what smells good, how should a blind test be designed, and when can one person's evaluation be trusted?