2026-11-07 · 8 min read

Stability testing at -25°C and 60°C: what a three-month run is looking for

Most factories test stability at 45°C and call it done. We run three months at -25°C and 60°C, because those are the two temperatures your carton will actually meet.

The short answer: -25°C simulates an unheated cargo hold or a winter warehouse; 60°C simulates a container on a Gulf dock in August. Testing at both ends for three months is how you find out whether the formula survives the journey, not whether it looks fine in a lab.

Why the middle of the range tells you nothing

A product stored at 25°C is a product nobody will ever buy. It is either in a container, in a warehouse, or in a bathroom — and all three are less friendly than a stability chamber set to room temperature.

The two conditions that actually break cosmetics in transit are cold and heat, and they break different things:

Cold attacks the emulsion. Water freezes, the oil phase separates, and on thawing the structure may not rebuild. A cream that looks fine at 25°C can arrive grainy.
Heat attacks everything else. Fragrance oxidises, actives degrade, sunscreen filters recrystallise, and packaging deforms or leaks.

Testing only the hot end is the common shortcut. It catches the dramatic failures and misses the slow ones that show up in your customer's second month.

What -25°C is standing in for

Cargo holds on long-haul freighters are not climate controlled, and neither are unheated warehouses in northern China, Europe or North America in January. -25°C is the realistic worst case, not an arbitrary number.

The specific failure to watch for is freeze-thaw cycling: water in the external phase crystallises, the emulsifier film is disrupted, and the system may separate on warming. One cycle often survives. Repeated cycles during a two-week journey are what matter.

What 60°C is standing in for

A shipping container sitting on a dock in Jeddah or Dubai in August is an oven. Internal temperatures can run well above ambient, and the cartons nearest the roof take the worst of it.

At 60°C the failures are different and faster:

Sunscreen filters, particularly organic UVB absorbers, can recrystallise out of solution
Fragrance components oxidise and the top note turns
Thickeners lose structure and the product thins or bleeds
Pumps and tubes soften; some leak
Preservative efficacy can drift as the system degrades around it

Three months, and what is checked

Samples go in at both temperatures and are pulled at intervals across three months. At each pull we look at:

Appearance — separation, syneresis, graininess, colour shift
Odour — oxidation is often smellable before it is measurable
pH — drift signals an unstable system or an interaction with packaging
Viscosity — thinning or thickening against the approved sample
Packaging — deformation, leakage, print adhesion, pump function

Anything that fails at either extreme is a formula problem, not a shipping problem, and it is much cheaper to find it here than in a customer complaint.

What passing stability does not prove

This is the part buyers usually are not told:

It does not prove a three-year shelf life. Accelerated testing gives an indication. A real shelf-life claim needs long-term data at ambient conditions, which takes years.
It does not prove compatibility with your specific pack. A formula can pass in a glass jar and fail in a pump. Pack compatibility is a separate test.
It does not prove microbiological safety. That is the preservative challenge test, which is a different protocol entirely.

When a supplier says "stability passed, three-year shelf life", ask which of the three above they actually tested. Usually it is one.

FAQ

How long does stability testing take?
Three months at both temperature extremes. It runs in parallel with your packaging and artwork work, not after it.

Can it be shortened?
For a first order, no. A shortened run gives a shortened answer, and the answer is the point.

Do all products need both extremes?
Products shipping to the Gulf need the hot end without exception. The cold end matters wherever the parcel will travel unheated, which is most of the world.

What happens if it fails?
The formula is adjusted and the clock restarts. It is better to lose a month here than a container later.