The observation
In an eight-hour comparison on a cooling patch, we measured mean skin temperature drop at two use levels for three cooling agents. At 0.2% all three were within 0.25 °C of each other. At 0.5% two of them lost between 0.6 and 0.8 °C of mean performance — while the ester held its level.
| Cooling agent | 0.2% w/w | 0.5% w/w | Change |
|---|---|---|---|
| Menthol citrate ester | 6.35 °C | 6.20 °C | −0.15 |
| Synthetic coolant (reference) | 6.10 °C | 5.50 °C | −0.60 |
| L-Menthol | 6.20 °C | 5.40 °C | −0.80 |
Internal comparative study, 2026. Protocol and raw data are released to qualified formulators under NDA. The figure is a mean over eight hours, not a peak.
Mechanism 1 — the solubility ceiling, and what happens above it
L-Menthol dissolves in water at roughly 0.4 g/L at room temperature. A patch matrix, a clear spray or a shampoo base all have a finite capacity for a lipophilic sensate, and that capacity falls as temperature drops and as surfactant and polymer compete for the same space.
Push past it and the excess does not become more cooling. It becomes crystals at the surface, an oily film, or a cloud point in a product that was meant to be clear. None of those deliver a better sensory result, and all of them are visible to a consumer.
Mechanism 2 — what leaves the bottle
Menthol has a boiling point around 216 °C but a high vapour pressure at skin temperature. In an open patch, a sachet with a poor barrier, or a spray bottle with headspace, the fraction that volatilises before use is not available on the skin. A higher initial load can simply mean a higher loss rate, and a shorter sensation.
This is the reason the ester exists: esterifying one hydroxyl with citric acid moves the boiling range to 320–350 °C and cuts volatility by more than 90%. The molecule stays where you put it.
Mechanism 3 — the formula binds it before the skin does
Anionic surfactants form micelles. Hydrophobic sensates partition into them, and a sensate inside a micelle is not available to a receptor. Add more sensate and you also push the system closer to cloud point, so the formulator adds more solubiliser or more surfactant — which binds more of the active. This is the loop that makes “more” produce “less” in a shampoo.
Mechanism 4 — the receptor stops listening
Cooling is mediated mainly by TRPM8. Like most ion channels, it desensitises under sustained or repeated high-level stimulation. A very high dose produces a sharp initial hit, then a flattened response — which a panel reports as “less cooling” even though the measured temperature drop in the first minute was larger. If your claim is about duration, dose is the wrong variable to optimise.
Finding your own ceiling
A dose ladder costs a week and saves a reformulation. What we run:
- Four levels — 0.1, 0.3, 0.5 and 0.8% — in the real base, not in water.
- Measure at 0, 2, 4 and 8 hours, with the base alone as a control.
- Store at 4 °C, 25 °C and 40 °C for four weeks, and inspect for crystals under polarised light microscopy.
- Weigh an open container over seven days to see how much leaves the product.
- Only then run the sensory panel, and score cool, burn, tingle and after-feel separately. They do not move together.
Four rules that come out of this
- Choose a grade for the phase you are in. Free acid for oil phases and emulsions; the 10% sodium salt solution for clear aqueous systems without a solubiliser.
- Add cold. Into the cooled phase, after the thickener. Heat is the fastest way to hydrolyse an ester back to menthol and citric acid.
- Do not chase intensity with dose. If 0.5% is not enough, the problem is usually availability, not quantity.
- Test duration, not just peak. The number a consumer remembers is how long it lasted.
Figures are from an internal comparative study on cooling patches; the full protocol is available under NDA.