Starting out · 3
Why 0.1% drowns out 30%: threshold decides who speaks in a formula
· 14 min read
The same single drop disappears in one material and dominates the bottle in another. The difference isn't how much you used — it's odour threshold. Using the 802 materials in our data that carry a recommended evaluation strength.
There's a failure every beginner meets: you weigh out a formula exactly to your ratios, and the result smells of one thing. And that one thing turns out to be the material with the smallest share in the blend.
The culprit is odour threshold, the concept beginners understand last and get punished by first.
What threshold means
Odour threshold is how dilute a molecule can get in air and still be detected.
What matters is the size of the gap. Thresholds vary between materials by orders of magnitude. Some materials need parts per million before you notice them; others are obvious at parts per billion.
So the weight ratio in your formula and the ratio of who is actually speaking are two different things. Most of what follows is the fallout from that.
What our data shows
Of the materials we've compiled, 802 carry a recommended evaluation strength. Grouped by that strength:
| Recommended strength for smelling | Materials |
|---|---|
| 10% or less | 527 |
| 1% or less | 204 |
| 0.1% or less | 45 |
Even for the simple act of smelling, the loudest group sits a hundredfold away from the mildest. And that's just evaluation strength; use levels in a formula are another question again.
So who are the 45? A few names:
2,3-dimethyl pyrazine(nutty)2-acetyl pyrazine(popcorn)2-acetyl thiazole(popcorn again)2,4-nonadienal(green)(E)-4-decenal(citrus)(R)-muscone(musk, which will come up again below)
Pyrazines and aldehydes dominate the list, which matches the literature: across chemical classes, aldehydes tend to have the lowest thresholds and alkanes the highest, with several orders of magnitude between them.
What this means for a formula
One: share is not presence. A 0.1%-class material may be the first thing you smell even at 0.5% of the blend. When something needs rebalancing, move that material first and leave the one sitting at 30% alone.
Two: make stock solutions first. Never weigh a low-threshold material straight into a formula; at the 0.1% level, a 0.01g error is fatal. Make a 1% or 0.1% dilution and weigh that instead. I learned this the annoying way, weighing neat material because dilution felt like a chore, and the finished blend smelled of exactly one thing. This is the other half of the dilution article: you dilute not only to smell accurately but to weigh accurately.
Three: not noticing a material doesn't mean it's doing nothing. Musks are the standard case. You may believe you can't smell muscone at all, yet remove it and the longevity and roundness of the whole blend collapse. Low-threshold materials keep working whether or not they reach your conscious attention.
Threshold is not the same as loudness
A common misreading: low threshold ≠ harsh smell.
Threshold describes how dilute you can go and still detect something; sharpness at full strength is a separate property. Plenty of low-threshold materials are extremely soft at working concentration. Musks again. Conversely, something that chokes you neat may not have a particularly low threshold. You're just smelling it neat.
To judge a material's role in a formula, take the threshold together with the strength you intend to use. Either number alone will mislead you.
Substantivity is a separate axis
Threshold decides who gets heard. Substantivity decides who is still there at the end. They're independent.
Among materials in our data that carry a substantivity figure:
| Materials | |
|---|---|
| Top (< 12 hours) | 421 |
| Heart (12–72 hours) | 595 |
| Base (> 72 hours) | 1,011 |
The longest is 3-heptyl dihydro-5-methyl-2(3H)-furanone at 901 hours, about 37 days.
That figure is persistence on a blotter, and skin is a different environment. But it makes one thing clear: top, heart and base run as a single race, every molecule off the line at the same moment, each evaporating at its own speed. The light material finishes first; the heavy material is still running.
The threshold ratio predicts dominance — but not infallibly
This article has argued that what decides who speaks in a formula is concentration divided by threshold. That ratio has a formal name in flavour research, the odour activity value (OAV), and its reliability has been tested directly.
In 2001 Grosch reviewed, in Chemical Senses, studies that rebuild real food aromas from synthetic blends of odorants: two wine varieties, three olive oils of different provenance, French fries, boiled beef and coffee. A model is built from analytical data until it matches the original aroma; then single components are omitted one at a time to see whether the aroma changes.
The findings come in two halves, and both matter. The first supports everything above:
The results indicate that odorants with higher odour activity values (OAV, the ratio of the concentration to the odour threshold) are frequently essential for the aroma.
The second is the caveat you need:
However, there are exceptions where odorants with high OAVs are suppressed in the aroma and compounds with lower OAVs are important contributors.
The paper discusses those exceptions in terms of the perceptual interactions between odorants in mixtures.
In practice: the threshold ratio is your first hypothesis, not your answer. It tells you which materials to suspect. To confirm that a material is actually contributing, there is exactly one method: take it out and smell again. You can run that omission test on your own three-material accords, and you should, often.
How to look this up
Every material page here carries strength and substantivity. You can also filter searches by substantivity directly; set a minimum of 72 hours and you have a shortlist of base note candidates.
For rigorous threshold figures, the standard reference is:
References
W. Grosch, Evaluation of the key odorants of foods by dilution experiments, aroma models and omission, Chemical Senses, 26(5), 533–545 (2001). PMID 11418500
L. J. van Gemert, Odour Thresholds: Compilations of Odour Threshold Values in Air, Water and Other Media, 2nd edition, Oliemans Punter, 2011, 486 pages, ISBN 978-90-810894-0-1.
It compiles roughly 17,000 threshold values from close to 2,930 references, and is the most complete compilation in the field — the kind of source worth citing when you're designing seriously or arguing specifications with a supplier.
Next up is formula structure: once you have ten materials, how to build a first accord, and why three materials teach you more than ten.