Yield, Plasticity, Strength, and Ductility

The turning point: where stretch becomes bend

In the last guide the bar behaved like a crowd of tiny springs. Pull gently and every atomic bond stretches a hair; let go and it snaps back exactly, obeying Hooke's law along a straight line whose slope is the Young's modulus. That is elastic deformation, and it is fully reversible. Now keep pulling. At some stress the bar stops coming all the way back — remove the load and a permanent stretch remains. You have crossed from elastic into plastic behaviour, and the material will never forget it.

On the atomic scale the two regimes are utterly different. Elastic strain just stretches bonds without changing who neighbours whom; plastic strain is slip — whole planes of atoms glide over one another as dislocations sweep through, and each atom ends up beside a new partner. That is why elastic strain vanishes and plastic strain stays. The stress that marks this hand-off is the yield strength, and it is the single number an engineer checks first, because it is the boundary between 'bends and recovers' and 'bends and stays bent.'

1. Trouble: for most metals the elastic line curves smoothly into the plastic region — there is no sharp knee, so you cannot just point at 'the' yield stress.
2. The fix (0.2 percent offset): on the strain axis mark a small permanent strain of 0.002, that is 0.2 percent.
3. From that mark draw a line parallel to the elastic slope (the same Young's modulus slope from the start of the curve).
4. Where that offset line crosses the stress-strain curve, read across to the stress axis: that value is the 0.2 percent offset yield strength (often called the proof stress). It is the stress that leaves exactly 0.2 percent permanent set behind.

Past yield: work hardening, the peak, and the neck

Once it yields, the metal does not just flow freely — the curve keeps climbing, because it takes more stress to push the deformation further. This is work hardening (strain hardening): as you deform the metal its dislocation density shoots up, and the swarming dislocations tangle and jam one another. It is exactly the paperclip you kink back and forth until the bent spot goes stiff and fights you. The metal is toughening itself up as you shape it.

STRESS (engineering = load / ORIGINAL area)
  ^
  |                 UTS
  |                 .-.
  |              .-'   `-._             necking:
  |            .'          `-.__  x  <- true area shrinks,
  |  YS ...--'                 fracture   eng. stress FALLS
  |        /:
  |       / :
  |      /  :  <- plastic flow = slip; rising curve
  |     /   :     is work hardening (dislocations jam)
  |  E /    :
  |slope    :
  | /       :
  +-+-------+--------------------------> STRAIN
   0     0.002
         offset line (parallel to E slope)
  |<----->| elastic: springs back, Hooke's law

The peak of that curve is the ultimate tensile strength (UTS), the largest engineering stress the bar carries. Just past it something visible happens: the deformation stops spreading evenly and localises into one narrowing waist — the neck. All further stretching now happens in that shrinking throat, so the load the bar can hold starts to drop, and the engineering curve turns downward toward fracture.

But here is the honest twist: the metal is not getting weaker after the UTS. The engineering stress divides the load by the original area, so as the neck shrinks that formula understates the real stress on the actual, smaller cross-section. If you instead divide by the true, shrinking area you get the true stress, which keeps rising all the way to fracture. In the uniform stretch before necking the two are linked simply: true stress = engineering stress times (1 + engineering strain), and true strain = ln(1 + engineering strain). The downturn of the engineering curve is a bookkeeping artifact of measuring against the old area — necking, not softening, is the real event.

Ductility: how far it bends before it breaks

Yield strength and UTS tell you how hard you can push; ductility tells you how far the material will stretch plastically before it finally parts. It is the length of the plastic road, and we measure it two ways. Percent elongation compares a marked gauge length before and after: %EL = (final length minus original length) / original length, times 100. Reduction of area compares the cross-section at the neck: %RA = (original area minus fracture area) / original area, times 100.

A quick worked example. A round bar starts with a 50 mm gauge length and a 12.5 mm diameter (original area = pi/4 times 12.5^2 = 122.7 mm^2). It fails at a gauge length of 64 mm and a necked diameter of 8.8 mm (fracture area = pi/4 times 8.8^2 = 60.8 mm^2). Then %EL = (64 minus 50)/50 times 100 = 28 percent, and %RA = (122.7 minus 60.8)/122.7 times 100 = 50 percent — a thoroughly ductile metal. One honest caveat: because most of the elongation happens in the tiny necked region, %EL depends on the gauge length you chose, so a ductility figure is only meaningful with its gauge length quoted (50 mm is the usual standard). %RA does not have that problem.

Ductility is really a spectrum. A ductile metal — copper, mild steel, most aluminium — necks and stretches visibly, giving you plenty of warning through ductile fracture, and it tolerates a scratch or bolt-hole because it can flow to spread the stress. A brittle material — grey cast iron, glass, a ceramic — runs almost straight up and snaps with barely any plastic strain and no neck at all, a brittle fracture that arrives with no warning. That difference is exactly why bridges and pressure vessels are built from ductile steel: you want a part that groans and bends before it fails, not one that shatters like a dropped mug.

Three different questions: stiff, strong, tough

The commonest beginner tangle is treating stiff, strong, and tough as one idea. They are three independent axes. Stiffness is the Young's modulus — resistance to elastic stretch, how much stress per unit of springy strain. Strength is a stress — the yield strength or UTS, how hard you can load it before it yields or breaks. Toughness is an energy — the total area under the whole stress-strain curve, how much work it soaks up before it fractures. A material can score high on one and low on another, and telling them apart is half of materials engineering.

Toughness needs both strength and ductility, because area = height times width, and you need a curve that is both tall (strong) and wide (ductile). This is why a single big number can mislead. Alumina ceramic is enormously strong yet its curve is a thin brittle spike, so its area — its toughness — is tiny; annealed copper is only modestly strong but stretches for miles, so its area is huge — tough but soft. A close cousin is resilience, the area under just the elastic part: the elastic energy a spring stores and gives back, whose modulus of resilience is (yield strength)^2 / (2 times E). For that 1500 MPa spring steel, that is roughly (1500 times 10^6)^2 / (2 times 200 times 10^9), about 5.6 MJ/m^3 of recoverable energy — which is precisely a spring's job.

And here is the law that governs the whole trade: almost every trick that raises strength does it by pinning dislocations, and pinning dislocations also robs the metal of ductility. Cold work, alloying, precipitates, and grinding the grain size finer all push yield strength up and stretch-to-break down together. That is the strength–ductility trade-off: 'stronger' almost always buys you 'less warning before it snaps.' The engineer's real skill is not maximising strength but choosing where on that trade-off a given part should sit.

From a datasheet number to a part you can trust

One last honesty. None of these numbers is a fixed constant — they scatter. Pull ten nominally identical bars and you get ten slightly different yield strengths, and for brittle materials the scatter is wild, because failure begins at the single worst flaw in the piece, so a big sample is likelier to hide a bad one (that is the Weibull story you meet in the fracture rung). Because of scatter, no responsible engineer designs a part to work right at its yield strength. Instead they divide the strength by a factor of safety to set a lower design (working) stress the part is actually allowed to see — a deliberate margin against the bad day, the unlucky bar, and the overload nobody predicted.

There is also a fast, cheap shortcut to strength worth flagging now. Instead of destroying a whole tensile bar, you can press a hard indenter into a surface and measure the dent — that is hardness, measured on the Brinell, Rockwell, or Vickers scales, and it is nearly non-destructive. For steels it tracks the UTS surprisingly well: a rough rule of thumb is UTS in MPa is about 3.45 times the Brinell number, so a steel at 200 HB has a UTS near 690 MPa. Two honest caveats: hardness is a correlation with strength, not a law, and — a classic trap — hardness is not the same as hardenability, which is about how deeply a steel hardens when you quench it. The next guide gives hardness, toughness, and safety factors the full treatment they deserve.

Step back and the plastic half of the curve has handed you a working vocabulary: yield strength (when it stops springing back), UTS (the peak load it bears), ductility (how far it stretches first), toughness (the energy it eats before breaking), and the trade-off that ties strength and ductility together. Keep them in separate boxes from stiffness, remember every number scatters, and you can read a materials datasheet the way an engineer does — not as a promise, but as a starting point to divide down into a stress the part will actually survive.