Silicone Heat Ageing: What Fails, and How Fast

What heat actually does to silicone

The siloxane backbone is one of the most thermally stable polymer chains there is, which is why silicone survives temperatures that would destroy any organic rubber. The weakness is not the backbone but everything around it: the crosslinks, the filler treatment, the additives and the residual volatiles.

At moderate temperatures the material continues to crosslink slowly, which makes it harder and less elastic. At higher temperatures the side groups on the chain begin to oxidise, which makes the surface chalky and eventually brittle. Above about 250 C the filler starts to affect the network and the material loses cohesion.

The practical consequence is that a silicone part rarely fails by melting or by suddenly losing strength. It fails by becoming stiff, losing its compression set resistance and then cracking at a stress concentration. The end of life is a mechanical failure caused by a chemical change.

That is why heat ageing is measured by retention of tensile strength and elongation rather than by appearance. A part that looks fine can have lost half its elongation and be unable to survive a bend that it took easily when new.

The retention curve, and the 25 degree rule

The chart on the left shows tensile strength retained against ageing time for the same compound at 200, 225 and 250 C. At 200 C it retains 78 percent after 336 hours and 66 percent after 672. At 250 C it is down to 38 percent at 336 hours and 22 at 672.

The rule of thumb is that the reaction rate doubles for every 10 to 15 C in organic polymers, and for silicone the effective interval is closer to 20 to 25 C. The curves in the chart are consistent with about 25 C: 672 hours at 200 C gives the same retention as roughly 170 hours at 225 C and 50 hours at 250 C.

That rule is what makes a short test useful. Seventy two hours at 225 C is a week of laboratory time and it predicts years of service at 150 C, within a factor of two or three. It is not a precise prediction, and it should not be presented as one, but it ranks compounds reliably.

The 70 percent line on the chart is the conventional limit for a part that has to keep working. A compound that crosses it at 336 hours at 250 C is not suitable for continuous service at 250 C, whatever its data sheet says about maximum temperature.

The five mechanisms

Further crosslinking at 32 percent is the largest single mechanism. It makes the material harder, reduces elongation and raises compression set. On a seal it is the mechanism that ends the life, because a harder seal takes less deflection and seals less well.

Loss of plasticiser and volatiles at 24 percent is the mechanism that shows first. A silicone that contains a plasticiser or a volatile process aid will lose it in the first hours at temperature, and the part shrinks and hardens. A post cured compound has already lost them, which is one reason the post cure matters.

Oxidation at 20 percent affects the surface before the bulk. Oxygen diffuses into the surface layer and attacks the side groups, which produces a chalky film. That film is a barrier, so the reaction slows, but the chalk is also a crack initiation site on a part that flexes.

Filler degradation at 14 percent and hydrolysis at 6 percent complete the list. Filler degradation matters at the top of the temperature range; hydrolysis matters in humid or steam service, where water attacks the filler-matrix bond.

Running the test properly

The standard test is ASTM D573: age the specimens in an air circulating oven at the specified temperature for the specified time, then test them at room temperature after conditioning. The air circulation matters, because a static oven builds up degradation products that change the result.

The oven has to be at temperature before the specimens go in, and the specimens have to be spaced so that air reaches all surfaces. A stack of specimens in a closed container is not a heat ageing test, it is a different experiment.

Test at least three times: the unaged control, the midpoint and the end point. Two points give a line; three give a check on whether the mechanism changed, which it often does. A curve that drops steeply and then flattens indicates a surface mechanism; one that accelerates indicates the network is failing.

Table 1 — Common ageing conditions and acceptance limits

Making a compound last longer

The first lever is the base polymer. A methyl vinyl silicone with a high vinyl content crosslinks more densely and resists further crosslinking better. A phenyl silicone extends the upper temperature range by 20 to 30 C at a significant cost.

The second is the post cure. A proper secondary vulcanisation removes the volatiles that would otherwise be lost in service, and it completes the crosslinking so that the material does not continue to harden in the application. It is the cheapest improvement available.

The third is the filler. A well treated fumed silica with a high surface area resists degradation better than a cheap precipitated filler. The difference shows at 250 C after 168 hours, and it does not show at all in a short test at 200 C.

The fourth is heat stabilisers. Iron oxide and cerium oxide are used to scavenge the radicals that drive oxidation, and they extend life at the top of the range. They also change the colour, which is why high temperature grades are often red or brown.

Predicting service life without guessing

The method is straightforward: test at three temperatures, plot the time to a chosen retention limit against the reciprocal of absolute temperature, and extrapolate. If the three points fall on a straight line, the mechanism is the same at all three and the extrapolation is defensible. If they do not, the mechanism has changed and the extrapolation is not valid.

The extrapolation should never go more than about 30 C below the lowest test temperature. Beyond that the uncertainty is larger than the answer, and a life prediction of 20 years is not something a 72 hour test can support.

The practical approach in most specifications is simpler and more honest: state a test condition and a retention limit, and let the part be qualified against it. That does not predict service life, but it does ensure that two suppliers are being compared on the same basis.

Table 2 — What each change buys, and where it shows

The post cure is first because it is a process step that most specifications require anyway. A compound that fails a heat ageing test without a post cure has not been tested; it has been tested incorrectly.