Alloy strength

Why is an alloy usually stronger than either of the pure metals inside it?

Copper is soft. Tin is softer still, and you can mark it with a fingernail. Mix them and you get bronze, hard enough to hold a cutting edge and to name an age. Iron is soft too; carbon on its own is soot or graphite, a lubricant. Together they make steel.

Mixing two weak things and getting a strong one looks like arithmetic failing. So the first question is not "what does the additive do" but something more basic: what is it about a pure metal that makes it weak in the first place?

Ask what has to happen for a metal to bend permanently. Its atoms sit in a regular crystal lattice, and permanent deformation means one plane of atoms sliding over another. If that required every bond across the plane to break at once, the calculation is straightforward, and it gives a strength on the order of a hundred times what real metals actually withstand. Real copper gives way absurdly early. Something is letting the planes slide cheaply.

That something is a defect: a dislocation, a line along which the lattice has one extra half-plane of atoms wedged in. Only the bonds in a narrow strip there are strained. Apply a modest stress and that strip re-bonds one row over, shifting the extra half-plane sideways by one atom; repeat, and the dislocation travels across the crystal until it reaches a surface, leaving the plane displaced by one atomic step. Nothing broke wholesale. A whole plane slipped, one row at a time, for almost nothing.

So a pure metal is soft not despite its perfect crystal but because that crystal is so uniform that a dislocation meets nothing to detain it. And that reframes the problem entirely. You are not trying to make the bonds stronger. You are trying to make the journey harder.

Which suggests what an added element is for. Zinc atoms in copper sit where copper atoms would, but they are the wrong size, and each one warps the lattice around it into a stress field a passing dislocation must push through. The stress needed to keep it moving rises. That is solid-solution strengthening, and it is why brass is harder than copper.

Carbon in iron does something related but sharper. A carbon atom is far too small to replace an iron atom; it wedges into the gaps between them instead. An interstitial atom strains the lattice much more per atom than a substitutional one, which is why a very small amount does so much — structural steels typically hold around 0.2 percent carbon by weight, and anything above roughly 2 percent stops being steel and becomes cast iron.

Two further routes follow from the same principle. Work hardening: deform a metal and its dislocations multiply and tangle until they obstruct each other, which is why a paperclip bent repeatedly at one spot first stiffens and then snaps there. And grain refinement: a real metal is a mosaic of small crystals in different orientations, and a dislocation cannot easily cross into a neighbour whose lattice planes point elsewhere, so finer-grained metal is stronger.

Steel gets its most dramatic property by combining these. Heat it until the carbon dissolves freely, then quench it so fast the carbon has no time to move out. The trapped carbon distorts the iron lattice into a strained, needle-like structure — martensite — extremely hard, and brittle enough to be nearly useless. Tempering, a gentle reheat, relaxes some of that strain and trades hardness back for toughness.

Which exposes the cost. Every one of these mechanisms works by denying dislocations their mobility, and dislocation mobility is exactly how a metal yields instead of cracking. Hardness and ductility trade against one another almost as a matter of definition, and most of metallurgy is the management of that trade.

A few refinements, in ascending order of how often they are missed.

"Alloys are stronger" is a serviceable rule with real exceptions. Solder alloys exist to melt low; some alloys are chosen for corrosion resistance, castability, or thermal expansion; and adding the wrong element in the wrong amount embrittles rather than strengthens. What is general is the mechanism, not the outcome.

The strengthening is not chemical bonding in the ordinary sense. Nobody is making iron-carbon bonds that are stronger than iron-iron bonds; the carbon is an obstacle in a lattice, and the lattice's own bonding is essentially unchanged.

And this is a crystal-scale story, not a composite one. A steel bar is not steel because two materials share the load between them the way concrete and reinforcing bar do; it is one crystalline material whose internal defects have been arranged to obstruct each other. Same word, "strength", two entirely different levels of description.

• Why some alloys strengthen further over days at room temperature, as in the age-hardened aluminium that made metal aircraft possible.
• How the same reasoning explains fatigue: why a metal fails after many small loads that each seemed harmless.

Every term the collection defines is gathered in the glossary.