Physical ageing of plastics

Why does a plastic part that has sat untouched in a drawer for a year snap where the same part fresh from the mould would bend?

Two identical mouldings from the same shot of the same machine. One is tested that afternoon and bends. The other spends a year in a dark drawer and then shatters under a load the first absorbed.

Nothing entered the drawer and nothing left: no light, no meaningful oxygen exposure, no load, no heat. If a material only changes when something happens to it, we have a contradiction. So the assumption to give up is that a solid at rest is a solid at equilibrium.

Clear away the mechanisms that are not doing the work, since each is real in its own place and each would show itself differently. Weathering — ultraviolet light breaking chains in the presence of oxygen — needs light, and there was none. Plasticiser migration, which stiffens flexible PVC, would show as lost weight; chain scission, as reduced molecular weight. Suppose we measure and find the part weighs what it did and its chains are as long as they were: no chemistry has occurred, and whatever changed is physical.

What could be physical about an unchanging solid? Ask how it was made. The polymer left the mould as a melt: chains writhing past one another, packing steadily more efficiently as the temperature fell. That exploration has a rate, and it collapses precipitously on cooling. Cooling has a rate too, set by the mould. At some temperature the first falls below the second and the chains can no longer finish rearranging before the temperature drops again. That crossing is the glass transition, and the solid below it is a glass.

Hold the consequence carefully, because everything follows from it. The glass is frozen in an arrangement appropriate to a warmer material, carrying more volume, more enthalpy and more disorder than equilibrium demands at its own temperature. It is not merely disordered; it is out of equilibrium.

So what does it do? What it was doing when it fell behind: pack. Slowly, locally, without any external cause, the chains keep shuffling into denser arrangements. That is physical ageing, and the drawer was never a place where nothing happened.

Now the timing, the strangest and most useful part. As the material densifies, the room the chains need in order to move disappears, so densification slows itself. The measured result, from Struik's long experiments, is that properties shift roughly linearly with the logarithm of elapsed time. Read that as it deserves: the change over one day, ten days, a hundred and a thousand is about the same in each step, so the first hours matter as much as the following year — and no laboratory can shorten the wait by patience, since another equal increment costs ten times as long again.

But why should denser mean more brittle? Two thresholds compete. Loaded hard enough, a glassy polymer either yields in shear — flowing locally, blunting stress, absorbing energy — or crazes and cracks, and whichever threshold the stress crosses first decides the character of the failure. Ageing raises the stress needed to yield, because yielding requires exactly the local mobility ageing has removed; the stress needed to open a craze rises much less. Push those two lines together long enough and they cross, and the part that used to bend now breaks — not because it got weaker but because it got harder to yield. Strength and durability part company, exactly as in a work-hardened metal.

One question settles the mechanism. If the glass really is out of equilibrium, thermodynamics is pushing it toward the denser state the whole time, at every temperature. Then why does a polymer stored far below its transition seem never to age, while one stored just under it ages in days? Because thermodynamics only names the direction. The rate is set by the segmental mobility the storage temperature allows, and that falls off a cliff below the transition — which is why shelf behaviour depends so sharply on the gap between room temperature and a material's own transition. PET's sits around seventy to eighty degrees Celsius, though quoted values shift with measurement method and sample moisture, so treat that as a marker rather than a constant.

Three refinements, the last of them a test.

Physical ageing is not degradation, and treating them as one is the common error. Degradation destroys molecules and cannot be undone; ageing rearranges nothing but packing and can be undone completely. The neighbouring case of a decaying film reel is its opposite in temperament too: acid released by the decay catalyses more decay, so it accelerates itself, while physical ageing removes the mobility it needs and decelerates itself.

The picture applies cleanly to amorphous glassy polymers. In semi-crystalline plastics a slow secondary crystallisation runs alongside it, and separating the two takes more than a hardness measurement.

And the falsification test is unusually decisive. Take the aged part, heat it briefly above its glass transition, and cool it at the rate the mould once did. If the account is right the ageing must be entirely erased — this is rejuvenation — and the part must recover its original impact behaviour with no change in mass or molecular weight. A calorimeter scan gives a quantitative version: an aged glass shows an endothermic overshoot as it passes the transition, its size tracking the enthalpy it had shed, and rejuvenation must remove that too. If annealing failed to restore ductility, or the sample had lost mass, or its chains proved shorter, chemistry was doing the work and this explanation is simply wrong.

• How annealing just below the transition stabilises precision mouldings, and what it costs in impact resistance.

Every term the collection defines is gathered in the glossary.