Why Things Break
Stress, Strain, and the Limits of Solid Things
Everything Is a Spring
A steel beam does not look springy. Load it and it is. Every solid object stretches when pulled, squashes when pressed, and bends when leaned on, and for small loads it does so in exact proportion – twice the load, twice the deflection – then springs back completely when released.
The reason is that a solid is atoms held at particular distances by their bonds. Pull the object and you are stretching billions of tiny bonds slightly, each of which pulls back harder the further it is taken from its preferred length. That is what a spring is. A steel bar is simply a very stiff one, and the floor you are standing on is sagging under you right now by an amount too small to see.
Every building is designed around this. The deflection is calculated, allowed for, and kept within limits that keep doors closing and glass unbroken. Structures are not rigid. They are stiff enough.
Thickness Decides Strength, Length Decides Stretch
Two ropes of the same material, one twice as thick as the other, will hold different loads. Two ropes of the same thickness but different lengths will hold the same load and stretch by different amounts. Separating those two statements is most of what materials science asks you to keep straight.
What a material actually cares about is two ratios. The first is force spread over cross-sectional area, called stress. The second is the fraction by which the object has been stretched, called strain. A meter-long wire stretched by one millimeter and a hundred-meter cable stretched by ten centimeters are in identical condition. Both are one part in a thousand longer than they were.
Working in those terms strips away the size of the object and leaves a property of the substance. Steel fails at a particular stress whether it is a pin or a bridge cable. That is what makes it possible to test a small sample and design a large structure.
When the Shape Stops Coming Back
Keep increasing the load and the neat proportionality eventually stops. Past a certain stress the material yields: it keeps deforming, and when the load comes off it does not go all the way back. Some of the shape change is permanent.
This is why a paperclip stays bent. Bend it gently and it springs back; bend it past its yield point and it holds the new shape. Every metal object that has ever been pressed, rolled, forged, or stamped exists because of this behavior, and every one that has ever been dented is a victim of it.
For a structural engineer, yielding is a friend. A steel beam that is overloaded sags visibly and audibly long before it lets go, and that warning is designed in. The load path also rearranges itself: as one part yields, it stops taking more and passes the excess to its neighbors, which is why a steel frame can survive damage that its arithmetic says should have finished it.
Two Very Different Ways to Fail
Some materials do all of the above. Others skip it entirely. Glass, cast iron, concrete in tension, and most ceramics have essentially no yielding stage. They stretch elastically right up to the moment they come apart, and then they come apart instantly.
The distinction is not academic. A ductile failure is loud, slow, and visible; a brittle one is sudden and total. Structural engineering has an enormous preference for the first, which is why steel is everywhere and why concrete is almost always reinforced with steel bars that take the pulling loads concrete cannot survive.
Worse, materials can switch. Steel that is comfortably ductile at room temperature can become brittle when cold, and several early welded ships broke in half in freezing water for exactly this reason. The material had not changed. The temperature had.
Everything Is Weaker Than It Should Be
Work out how strong a material ought to be from the strength of its atomic bonds and you get a number. Measure a real sample and you get something between a hundred and a thousand times smaller. Almost every solid object around you is failing to be strong by two or three orders of magnitude, and the reason is one of the more satisfying results in engineering.
Real materials contain flaws – scratches, voids, inclusions, microscopic cracks. Load flowing through the material has to detour around them, and it concentrates savagely at the tip. A crack that is sharp enough can multiply the local stress a hundredfold, so the material at that one point reaches its breaking stress while the object as a whole is nowhere near it. It fails there, the crack advances, the tip stays sharp, and the failure runs.
Griffith demonstrated this around 1920 with glass fibers, finding that thinner fibers were dramatically stronger – because a thinner fiber simply has less surface available to hold a big flaw. A glass cutter does not cut glass. It puts a controlled scratch on the surface, and the glass then obligingly breaks along it under a gentle bend.
Nearly every strength trick in engineering is a way of dealing with cracks. Drilling a round hole at the end of a crack blunts the tip and can stop it. Toughened glass is cooled so its surface ends up in compression, squeezing cracks shut. Composites interleave fibers so a crack running through one is stopped when it reaches the boundary of the next.
Bigger Is Weaker
Scale an object up without changing its shape and its strength does not keep pace with its weight. Strength depends on cross-sectional area, which grows with the square of the size. Weight depends on volume, which grows with the cube. Double everything and you have four times the strength holding up eight times the load.
This is why an ant can carry many times its own weight and an elephant cannot. It is why an elephant’s legs are proportionally far thicker columns than a mouse’s. Large animals have to change shape rather than simply scale up, and an insect enlarged to human size would collapse under itself before it managed a step.
Galileo worked this out and drew the consequence explicitly: there is a maximum size for a land animal built on a given plan, and for a tree, and for a building made of a given material. The rule also runs the other way, which is why very small things are so robust. An insect falls from any height and walks away, because its strength has not shrunk nearly as fast as its weight has.
What the Atoms Are Actually Doing
Everything above treats material as a smooth substance with properties like stiffness and yield stress. Those numbers are real and measurable, and none of them is fundamental. They are all consequences of what the atoms inside are doing.
Yielding in a metal is not bonds snapping in unison. It is the movement of line defects through the crystal lattice, which lets whole planes of atoms slip past each other one row at a time – enormously easier than breaking them all at once, which is exactly why real metals yield so far below their theoretical strength. Hardening a metal means making those defects harder to move, whether by alloying, cold working, or heat treatment.
The bonds themselves are electromagnetic, and their stiffness comes from quantum mechanics – specifically from the exclusion principle that stops electron clouds from overlapping. The reason a solid resists being squashed at all traces back to spin and the rule that no two electrons may occupy the same state.
Strength Is a Property of Flaws
Strength turns out not to be a property of a substance so much as a property of a substance plus its flaws. The perfect version of almost any material would be extraordinary. The real version is limited by the worst crack in it, which is why so much of engineering is not about making things stronger but about controlling, blunting, and designing around defects.
It is also a good reminder that the everyday scale has laws of its own that are not visible from below. Nothing about a single atomic bond tells you that a scratch on a windowpane matters, or that there is a largest possible tree. Those facts live at the scale where enormous numbers of atoms act together, and they are as real as anything underneath them.
