Accelerated ageing for balms: the protocols, the maths and the honest limits

Oven ageing, freeze thaw cycling and centrifuge screening explained, with the Q10 assumption examined, ASTM F1980 in context and the failures they never predict.

An accelerated ageing programme buys you two things: early warning that a formula will fail, and a documented reason for the date you print. It does not buy you a shelf life. This page gives the chamber conditions, the freeze thaw and centrifuge screens, the arithmetic that converts weeks in an oven into months on a shelf, and the specific failures where that arithmetic is wrong by a factor of three.

• Standard set: 4, 25, 40, 45 C
• 12 weeks minimum
• Q10 = 2 is an assumption
• ASTM F1980 is a packaging standard
• Real time confirms, always

What accelerated ageing is actually for

Accelerated testing does one job well and one badly. It is excellent at finding failures: put two candidate formulas at 40 C for six weeks and the one that sweats, separates, grains, fades or eats its own label will say so. That is a comparison under identical stress, valid without any modelling assumption at all. It is poor at producing dates, because a date requires a quantitative relationship between chamber and shelf, and for an anhydrous balm that relationship takes several different values depending on which failure mode you are watching.

Keep the two uses separate in your paperwork. "Formula B survived a condition that killed formula A" is a finding. "Twelve weeks at 40 C therefore equals twelve months at room temperature" is a model, only as good as the constant inside it. The end points worth measuring are set out under shelf life and stability testing.

The standard chamber set and how long to hold it

A conventional programme runs four temperatures in parallel, each on samples in the final container with the final closure and label, held for 12 weeks at minimum. Anything tested in a beaker tests the formula only, which misses roughly half of what goes wrong.

Relative humidity, specified as 60 or 75 percent in borrowed pharmaceutical protocols, is irrelevant to the formula and important to the pack: it drives label adhesive failure, board tube softening and tinplate corrosion, so it belongs in a packaging compatibility study. If humid chamber samples come out with rust spots on the tins or labels lifting at the edges, that is a real result about your packaging supplier.

The Q10 arithmetic, written out

Every accelerated-to-real-time conversion in cosmetics rests on the same expression, which is the Arrhenius relationship expressed as a temperature coefficient:

Acceleration factor = Q10 raised to the power of (T accelerated minus T ambient) divided by 10.

Q10 is the factor by which the rate rises per 10 degree increase. The default value of 2 corresponds to an activation energy of roughly 53 kJ/mol near 300 K, a plausible mid range figure for many degradation reactions and not a measurement of yours. Taking Q10 as 2 and ambient as 20 C, 40 C is two intervals hotter, so the factor is 4. Twelve weeks times 4 is 48 weeks, a shade over 11 months. That one line of arithmetic is the entire basis of the claim that three months at 40 C equals a year.

Now vary the two assumptions inside their honest ranges. Published discussion of Q10 for packaging and chemical ageing puts it between about 1.8 and 2.5, and "ambient" is quoted as 20 C in Europe and 25 C in most international guidance.

The same twelve weeks in the same oven supports anything from 6.7 to 17.3 months, a spread of two and a half times, without a single data point changing. The reference temperature is doing as much work as the protocol. So write the conclusion as "12 weeks at 40 C, no change against the frozen control, read as approximately 11 months on a Q10 of 2 against a 20 C reference", not "one year stability confirmed", and file it with your batch records.

Three places the Q10 model breaks

The expression assumes one reaction, with one activation energy, whose mechanism does not change across the temperature range. Balms break all three assumptions, in three different ways.

Lipid oxidation changes mechanism with temperature

Autoxidation is a chain of steps with different activation energies, and heat does not scale them equally. Above roughly 50 to 60 C, hydroperoxide decomposition accelerates faster than hydroperoxide formation, so the primary oxidation pool shrinks while secondary aldehydes rise. Oxygen solubility in oil also falls as temperature rises, so a hot sample can become oxygen limited in a way an ambient one never is. A hot oven does not run the ambient reaction faster, it runs a differently balanced one. The mechanism is under rancidity and oxidation.

Crystallisation is not a rate you can accelerate

Grain is not a reaction. It is a phase transformation plus Ostwald ripening, with an optimum temperature band rather than a monotonic rate. Shea stearin held near 20 C works through beta prime toward beta over days to weeks. Held at 40 C, above the melting range of much of the crystal population, that transformation has no solid phase to happen in, and the sample recrystallises from a fresh melt when it cools with no memory of the twelve weeks. A 40 C oven can therefore return a clean pass on a formula that grains reliably in a customer's drawer. The mechanism and the fault are both set out on grainy shea butter.

The correct accelerator for grain is cycling, not a hot hold. Repeated passes through the partial melting range drive ripening far harder than a steady temperature at either end, which is what a product experiences in a bag, a car or a container. It is also why balm sweating shows up in cycled samples that steady-state samples never reveal.

Fragrance loss is evaporation, not reaction

Scent drift is partly terpene oxidation but mostly loss of the volatile fraction through the closure into the headspace. That is governed by vapour pressure, which rises with temperature according to Clausius and Clapeyron at a different rate for every component in the blend. Heating a fragranced balm does not fade it uniformly, it reweights it: the light citrus top goes disproportionately, leaving a sample that smells different rather than weaker. A 40 C oven overstates top note loss and understates the slow oxidative turn that arrives months later. Both are covered under scent fades in balm.

ASTM F1980 is a packaging standard, not a cosmetic shelf life method

ASTM F1980, "Standard Guide for Accelerated Aging of Sterile Barrier Systems for Medical Devices", is the document almost every online accelerated ageing calculator is silently derived from, including the Q10 default of 2 and the practice of capping the chamber temperature to avoid changing the failure mechanism. It is quoted far outside its scope.

It is a voluntary consensus standard published by ASTM International, a standards body and not a regulator. Its subject is the sterile barrier system, meaning the seal and material of a medical device package, and the property predicted is seal integrity over time. It binds nobody except where a party adopts it or a regulator recognises it in a device submission. It has no jurisdiction over cosmetics anywhere, it does not address the chemistry of the product inside the pack, and it states in its own terms that accelerated ageing does not substitute for real time ageing, which remains the definitive evidence.

Freeze thaw cycling: five cycles, and what each one is asking

The harsher of the two common protocols runs five cycles of 24 hours at -10 C followed by 24 hours at 40 C, ten days in total, on units in final packaging. A gentler version runs three cycles between -10 C and 25 C, the variant used in the small maker protocol under shelf life and stability testing. Neither is a standard. Both are conventions, and the five-cycle version is the one for a product that will be posted, sold outdoors or shipped in an unheated van.

Each cycle asks four questions at once. Does the crystal network survive repeated partial melting, or does the surface come back rough, sweaty or grained? Does the closure survive the headspace air expanding and contracting on every leg? Does the label adhesive survive condensation on the way up from -10 C? And does oil pool at the surface or the wall? Photograph each unit at the same distance and light at the end of every thaw leg, because the difference between cycle one and cycle five is obvious in pictures and invisible in memory.

The centrifuge screen, and what 3,000 rpm actually means

The standard cosmetic separation screen is 3,000 rpm for 30 minutes, and it is the most loosely specified test in common use, for two reasons.

First, rpm is not a force. Relative centrifugal force depends on rotor radius: RCF equals 1.118 times ten to the minus five, times radius in centimetres, times rpm squared. At 3,000 rpm an 8 cm rotor delivers about 805 times gravity, 10 cm about 1,006, 15 cm about 1,509 and 20 cm about 2,012. Two labs both reporting "3,000 rpm, 30 minutes" can differ by a factor of two and a half in applied stress. Record the RCF, not the rpm.

Second, a set balm is a yield stress solid, so below a threshold stress nothing moves at all. Taking a density difference of about 80 kg per cubic metre between wax crystals and liquid oil, an acceleration of 1,000 times gravity and a 3 cm sample column, the driving stress works out near 24 kPa. That is an order of magnitude estimate from stated assumptions, not a measurement, but the conclusion it supports is robust: a soft body balm with a yield stress around 1 kPa will separate under that load, while a hard lip balm well above 24 kPa will not, and its clean result means only that it was too firm to test. Yield stress and how to think about it is covered under rheology and yield stress.

The screen earns its place in two situations: comparing soft, high-oil formulas against each other, and testing anything at the temperature where it is actually vulnerable, which usually means spinning warm samples at 40 to 50 C. Used that way it is a fast pass or fail on whether a blend holds its oil, and it complements the visual evidence under balm separated in layers. Used cold on a firm stick it is theatre.

The Schaal oven test at 63 C, and what it maps onto

The Schaal oven test is the oldest accelerated oxidation method still in routine use, borrowed from the edible oil trade. Samples sit in open beakers or in the intended pack in a forced-draught oven at 63 C, some sources use 65 C, assessed at intervals by trained sensory panel and by peroxide value. The trade convention is that one day at 63 C corresponds to about one month at ambient, the end point being the first reliably detected off-note or a peroxide value threshold such as 20 milliequivalents per kilogram.

Check that convention against the Q10 model. From 20 C to 63 C is 43 degrees, and a Q10 of 2 gives a factor of about 19.7, so one day maps to roughly 20 days rather than 30. Reproducing "one day equals one month" exactly requires a Q10 of about 2.2. Two independently derived conventions landing within 50 percent of each other is mild reassurance that the model is not absurd. It is not evidence that either is right for your product.

The practical limit is physical. At 63 C almost every balm is liquid, so a Schaal test on a finished balm assesses the oil phase and the headspace without the crystal network. That makes it a good way to rank oil blends before you formulate, the same job the oil blend calculator does on paper, and a poor way to judge a finished product. Run it on the blend, not the stick, and pair peroxide value with a secondary measure, because peroxide value falls again once hydroperoxides start decomposing.

The real time study you still have to run

Every framework that takes accelerated data seriously also requires real time data, and treats the real time arm as the one that decides. The pharmaceutical guideline ICH Q1E, which governs medicinal products for regulators in the EU, the US and Japan and has no force over cosmetics at all, contains a rule of thumb worth borrowing precisely because it is conservative: a shelf life may be extrapolated to no more than twice the period covered by long term data, and in no case more than 12 months beyond it.

Applied in a balm workshop, six months of real time data at 25 C supports a 12 month claim and twelve months supports a 24 month claim. Accelerated data supports the extrapolation, it does not replace the base. Under that rule a brand new formula cannot honestly carry a two year date in its first year of existence, however well it did in the oven. Claim 12 months, keep testing, lengthen the claim at the next label revision. The starting ceiling from the shortest lived oil in the blend comes off the shelf life calculator.

The real time arm costs almost nothing: two units per batch in final packaging, ambient and dark, plus a frozen control, compared monthly and written down whether or not anything changed. An unbroken run of "no change" entries is the only evidence that actually supports a date. The symptoms to watch for are on rancid balm and balm went off before its best before.

What the rules require, and what you have to decide yourself

No jurisdiction prescribes an accelerated ageing protocol for cosmetics. The instruments require an outcome and leave the method to you and to whoever signs your safety assessment.

In the European Union, Regulation (EC) No 1223/2009 requires a Cosmetic Product Safety Report, and Annex I Part A requires the stability of the product under reasonably foreseeable storage conditions to be addressed, together with the minimum durability date that follows from it. Great Britain applies the same requirement through the retained UK Cosmetics Regulation, and Northern Ireland continues under the EU instrument. Neither names a chamber, a temperature or a duration. The assessor decides whether your evidence is adequate, and the route through the safety report is set out under safety assessment and the CPSR and selling balms in the UK and EU.

In the United States, the Modernization of Cosmetics Regulation Act of 2022 requires adequate substantiation of safety for a cosmetic product, again without prescribing a stability method, and the good manufacturing practice regulations it directed the FDA to write were still working through the rulemaking process rather than in force at the time of writing. Check the FDA's current position rather than relying on any secondary summary, including this one. The duties as they stand are under MoCRA duties.

Three limits are worth stating plainly. Nobody can tell you what Q10 value applies to your formula, because it has not been measured and it differs by failure mode. Nobody here can tell you whether your protocol satisfies your safety assessor, because that is a professional judgement on your specific product by a qualified person. And no oven result converts into a legally defensible date, because no such conversion exists for lipid systems. What does hold: run the four chambers to find failures, run the cycling to find what a steady hold hides, state your Q10 and reference temperature every time you quote a converted figure, never claim more than twice your real time evidence, and treat every accelerated number as a hypothesis the retained samples are still testing.

Reviewed and updated 6 September 2026. Spotted an error? Tell us and we will fix and log it.