Yield stress, ultimate stress — material behavior: Visual walkthrough

Step 1 — Pull a rod and measure two honest numbers

WHAT. Take a metal rod. Grab both ends and pull with a steady force. It gets a tiny bit longer. We are going to record exactly two things as we pull harder and harder.

WHY. A raw force in newtons and a raw stretch in millimetres are unfair comparisons: a fat rod takes more force than a thin one to do the same thing, and a long rod stretches more than a short one for the same fractional change. To compare materials and not shapes, we must divide the force and the stretch by the rod's own size. That gives us two "fair", shape-free numbers.

PICTURE. Below: the rod, its original length $L_{0}$ (the length before any pulling), its cross-section area $A_{0}$ (the area of the flat end you'd see if you sawed it — the red slice), the pulling force $F$ , and the tiny extra length $Δ L$ it gains.

Figure — Yield stress, ultimate stress — material behavior

The whole rest of this page is a single graph: stress $σ$ going up, strain $ε$ going right. Each time we pull harder we plot one dot. Let us start plotting.

Step 2 — The first dots come out in a straight line

WHAT. For small pulls, plot $σ$ (up) against $ε$ (right). The dots fall on a perfectly straight line through the origin.

WHY. Zoom into the metal: atoms are held to their neighbours by bonds that behave like tiny springs. Stretch a spring a little and the pull-back force grows in proportion to how far you stretched it. Add up millions of these tiny linear springs across the cross-section and the whole rod also responds linearly: double the strain, double the stress. Straight line.

PICTURE. The red segment is the data so far — dead straight, aimed at the origin. Beneath it, the atomic springs stretched by a small even amount.

Step 3 — The line stops being straight (the proportional limit)

WHAT. Keep pulling. At one particular dot the data stops hugging the straight line and starts to bend over to the right.

WHY. The atomic-spring picture only stays linear while each bond is stretched gently. Push the bonds far enough and their restoring force is no longer simply proportional to the stretch — the "spring" softens. The last dot where $σ = E ε$ still holds exactly is called the proportional limit. Past it, our clean straight-line formula is no longer the truth.

PICTURE. The straight dashed line is Hooke's law extended (what would happen if the springs stayed linear). The red curve is the real data peeling off it. The little circle marks the last honest straight-line point.

Recall Is the rod still elastic here?

Just past the proportional limit ::: yes, for a whisker it still returns on unloading — it is curved but not yet permanent. The straightness and the returning are two slightly different milestones; we keep them separate on purpose.

Step 4 — It stops springing back: yield

WHAT. Pull a little more. Now unload — and the rod does not return to $L_{0}$ . It is a bit longer forever. The stress where this permanence sets in is the yield stress $σ_{y}$ .

WHY. At the atomic scale, whole planes of atoms suddenly slip over one another and re-bond in new positions — a defect called a dislocation glides through the crystal. Slipping and re-bonding is a one-way trip: the atoms don't march back when you release the load. So part of the strain is now plastic (permanent) rather than elastic (recoverable).

PICTURE. Two rows of atoms; the red plane has slipped one step to the right and stayed there. On the graph, the point $σ_{y}$ where recoverable behaviour ends.

Step 5 — The gradual-yield problem, and the 0.2 % offset fix

WHAT. For many spacecraft alloys the curve does not have a sharp kink at yield — it bends over so smoothly that two engineers eyeballing it would pick different points. We need a rule that gives everyone the same number.

WHY. "Yield" physically means "a definite amount of permanent strain has occurred." So we pick a definite amount — 0.002, i.e. 0.2 % permanent strain — and define yield as the stress that leaves exactly that much behind. To find it on the graph: draw a line with the elastic slope $E$ but shifted to the right so it starts at $ε = 0.002$ . Where that line stabs the real curve is $σ_{y}$ , by definition. It is repeatable because $E$ and $0.002$ are fixed for everyone.

PICTURE. The dashed offset line is parallel to the elastic slope, shifted right by 0.002. The red dot where it crosses the curve is $σ_{y}$ . The horizontal gap marks the locked-in 0.2 % permanent strain.

Step 6 — It fights back harder: strain hardening up to the peak

WHAT. Past yield the curve keeps climbing, but along a shallow, arching path — not the steep elastic line. It rises to a single highest point, the ultimate tensile stress $σ_{u}$ .

WHY. Every slip spawns more dislocations, and they tangle and jam each other like too many cars merging. A jammed dislocation is harder to move, so it takes more stress to keep deforming — the material has effectively toughened itself. This is strain hardening. The stress can only keep rising until the tangling can no longer outpace the thinning we meet in Step 7 — that balance point is the peak $σ_{u}$ .

PICTURE. Red curve arching gently upward from $σ_{y}$ to its summit $σ_{u}$ ; a knot of tangled dislocation lines drawn below to show why it stiffens.

Step 7 — Necking, the falling tail, and fracture

WHAT. After the peak the engineering-stress curve goes DOWN even though the material is closer to breaking, until it snaps at the fracture point.

WHY. At $σ_{u}$ one spot starts thinning faster than the rest — a neck forms, like the narrow waist of a stretched piece of gum. From now on all the stretching crowds into that shrinking neck. The real area there is dropping fast, but our formula $σ = F / A_{0}$ still divides by the original area $A_{0}$ . Because $A_{0}$ can't shrink in our arithmetic, and the force $F$ now falls (there's less material carrying it), the engineering stress we plot slides downhill — an artifact of using $A_{0}$ , not a sign the metal got weaker. Follow it down and it fractures.

PICTURE. The red neck on the specimen; the engineering curve dropping to the fracture cross, with a dashed "true stress" curve (using the real shrinking area) still rising — to show the drop is a bookkeeping effect of $A_{0}$ .

Step 8 — Degenerate and edge cases (never get surprised)

WHAT / WHY / PICTURE, four quick cases the reader must never trip over:

Recall Which corner is a spacecraft strut usually kept in?

Operating region ::: deep inside the straight elastic segment near the origin — we design so real stress stays below $σ_{y}$ with margin, so we never leave Step 2's straight line. See Factor of safety and margins and Structural load cases and launch loads.

The one-picture summary

Here is the whole derivation compressed into a single annotated curve — origin, straight elastic climb, proportional limit, the offset-defined $σ_{y}$ , the strain-hardening arch to $σ_{u}$ , and the necking drop to fracture. Every landmark you built, in order.

Recall Feynman: the whole walkthrough in plain words

We pulled a metal rod and, to be fair to every shape of rod, we tracked two ratios instead of raw numbers: stress (pull ÷ original end-area) and strain (stretch ÷ original length). Plotting stress up against strain right, the first dots came out in a straight line — because the atomic bonds act like tiny springs, and springs pull back in proportion to their stretch; that line's steepness is Young's modulus $E$ . Pull past a point and the line bends (proportional limit), then reaches yield — where planes of atoms slip for good and the rod stays permanently longer. Because that bend is often too smooth to read, we invented the 0.2 % offset: shift the elastic line right by 0.002 and where it hits the curve is yield. Beyond yield the tangled slips make the metal fight harder (strain hardening) up to the peak, the ultimate stress. Then one spot necks, our $F / A_{0}$ formula makes the plotted curve slide down, and the rod fractures. Glassy stuff skips the middle and just snaps; soft metal has a long plateau; push instead of pull and the whole picture mirrors into the negatives. Spacecraft live in that first straight stretch, comfortably below yield.

Flashcards

Why divide force by original area to get stress?

To make a shape-free, fair number that compares materials not fat rods,

σ = F / A_{0}

Why is the first part of the curve straight?

Atomic bonds act like linear springs, so stress is proportional to strain,

σ = E ε

What does the slope of the elastic line equal?

Young's modulus

E

— stress needed per unit strain (stiffness).

What physically happens at yield?

Dislocations glide — planes of atoms slip permanently, so strain no longer fully recovers.

Why is the 0.2% offset method needed?

Alloys yield gradually; fixing 0.2% permanent strain gives one repeatable yield stress for everyone.

Offset yield formula?

σ = E (ε - 0.002)

intersected with the real curve.

Why does the engineering curve drop after

σ_{u}

Necking shrinks the real area but we still divide by fixed

A_{0}

, so plotted stress falls.

Engineering vs true stress after necking?

Engineering (uses

A_{0}

) drops; true (uses shrinking real area) keeps rising.

Brittle vs ductile curve shape?

Brittle: nearly straight then snaps (

σ_{y} \approx σ_{u}

); ductile: long plastic plateau before fracture.

Under compression what happens to the elastic slope?

Same slope

E

, mirrored into negative

σ

and

ε