Plastics creep and metals do not

Slow deformation under constant load, and the parts it ruins.

Plastics creep and, under any load an ordinary design ever puts on them at everyday temperatures, metals do not, because a plastic's long molecular chains can slowly slide and rearrange past one another under a steady load that a metal's tightly bonded crystal lattice simply resists indefinitely at the same stress and the same temperature.

Eberhart's Why Things Break was not the book I expected when I first opened it in graduate school, since a title like that promises snapping and shattering, and instead a good part of it is spent on failures too slow to notice happening at all, a part quietly bending further and further away from its original shape for months before anyone thinks to measure it.

The mechanism behind creep

Every solid deforms a little the instant a load is applied, and in a well-designed metal part staying safely within its elastic range, that deformation simply stops there, holding steady indefinitely as long as the load itself does not change. Many plastics behave quite differently under the same kind of steady load, continuing to deform slowly and continuously for as long as the load remains applied, a phenomenon called creep, driven by the plastic's long tangled molecular chains gradually sliding and reorganising past one another under sustained stress rather than locking rigidly into place the way a metal's crystal structure does. The load never has to increase for creep to continue, since the mechanism is time-dependent rather than load-dependent, meaning a plastic part can keep slowly bending away from its original shape for weeks or months under a load that never itself grew any larger. Creep of this kind typically runs fastest in the first hours or days after a load is applied, slows to a steadier rate for a long stretch afterward, and in a part genuinely overloaded for the material can eventually accelerate again shortly before it fails outright.

The coat-hanger comparison

A plastic coat hanger left holding a heavy winter coat in a wardrobe for months can be found, come spring, hanging in a visibly permanent curve, even though the coat's weight never increased even slightly beyond what the hanger comfortably held on the very first day it was hung. A wire coat hanger of similar shape, loaded with the same weight for the same length of time, shows no equivalent sag at all, since a metal well within its elastic limit simply does not continue deforming under a load that stays constant, no matter how long that load is left in place. The plastic hanger was never overloaded in the ordinary sense, since nothing about the weight it carried ever changed; it was simply given enough time for a mechanism metal does not share to do its slow work. A plastic garden chair left assembled outdoors through a whole summer, its seat slowly deepening into a permanent dip under nothing heavier than the people who occasionally sit in it, tells the same story on a larger scale, sagging not from any overload but simply from carrying ordinary weight for month after month without a rest.

Why plastics are so much more prone to this than metals

The difference comes down to how each material's internal structure responds to time itself, not only to load. A metal's atoms sit in a tightly bonded, repeating crystal lattice that resists rearranging itself at ordinary temperatures, so a load well within the elastic range produces a deformation that is essentially permanent in its smallness, not growing further no matter how long the load is sustained. A plastic's structure is built instead from long tangled chains loosely bonded to one another, held in place more by friction and entanglement than by the strong fixed bonds a metal relies on, and under a sustained load those chains can slowly slip past each other, a little further with every passing hour, in a way no comparable slippage is available to a metal's rigid lattice at the same everyday temperature. Warmth speeds this slipping along considerably, since heat gives the tangled chains more energy to wriggle past one another, which is why a plastic part creeps distinctly faster on a hot day than a cold one under an identical load.

The number that matters here

A loaded plastic component can continue deforming for weeks or months under a load that is only a fraction of its rated strength, sometimes doubling its original deflection over that time even though the load itself never changed at all, a slow continuation that a strength calculation performed only for the instant the load is first applied has no way of predicting.

What this changes in practice

Designing a plastic part to carry a sustained load for any real length of time means checking its behaviour under that load held constant for the actual duration the part will experience, not only its behaviour the moment the load is first applied, since a plastic that comfortably passes an instant strength test can still be entirely unsuitable for carrying the same load for months on end. A metal part, in the same everyday temperature range, rarely needs this second check at all, which is exactly the habit of mind that catches designers out the first time a plastic replacement quietly fails a job a metal part had done for years without complaint. A prototype tested only briefly on a bench, loaded and unloaded within a single afternoon, will pass a strength check with real margin to spare and still say nothing at all about what the same part does after carrying that same load, unremarked, for the following six months.

Where this stops being true

Metals are not actually immune to creep in every circumstance, only within the ordinary temperature range most everyday designs operate in. Raised to a large enough fraction of their own melting point, a jet engine's turbine blades being the clearest working example, metals creep too, and by essentially the same underlying mechanism, atoms slowly rearranging under sustained stress once there is enough thermal energy available to let them move. The genuine distinction is not that metals cannot creep at all, it is that plastics reach the temperature range where creep becomes a real practical concern at ordinary room temperature, while most structural metals only reach their own equivalent range at temperatures a normal design never comes close to.