Corrosion
A Socratic walk-through of corrosion — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does iron rust while gold left in the same damp air does not?
Leave an iron nail and a gold ring on the same damp windowsill for a month. The nail is ruined and the ring is untouched, so the difference must live inside the metals. But then consider aluminium, which is more eager to react with oxygen than iron is, and which sits outdoors for decades looking fine. Eagerness alone clearly does not decide the outcome. What else is doing the deciding?
Reasoning it through
REASONING #Begin with what rusting is. Iron atoms give up electrons and leave the metal as ions. Where do the electrons go? Something must accept them, and in damp air that is dissolved oxygen, which takes them up together with water to form hydroxide. Two transactions, then, one releasing electrons and one consuming them, and neither runs without the other.
Here is the part that reframes everything: they need not happen in the same place. Electrons travel freely through metal, ions through a water film. So one patch of a nail can give up electrons while a patch some distance away is where oxygen takes them, the metal carrying current one way and the wet film the other. What have we just described? A short-circuited battery, built out of a single object. Which is why rusting needs water, and why salt makes it so much worse: salt does not attack the iron, it makes the film a better conductor and so completes the circuit.
That explains how iron rusts, not why gold does not. What does the electron-releasing step cost? For gold, a great deal — it clings to its electrons so tightly that ordinary oxygen and water cannot pay the price, so there is no reaction and no circuit. Gold's inertness is a transaction never worth making.
Aluminium is the opposite case, and more instructive. Its electron-releasing step is cheap; freshly cut aluminium reacts with air almost at once. So why does it stop? Because of what the product does. Aluminium oxide forms a thin, dense, tightly-bonded film occupying nearly the same volume as the metal it replaced. It seals the surface, oxygen and water can no longer reach fresh metal, and the reaction starves — and scratch it, the film reforms in moments.
Now put iron's product beside it. Rust is hydrated iron oxide, and it takes up roughly twice the volume of the metal consumed. What must a coating do when it is far bulkier than the surface it grew from? It cannot lie flat. It cracks, blisters, and flakes, exposing clean metal — and the process begins again there. That is the crux: iron rusts not because it is unusually reactive, but because its oxide never becomes a barrier.
The analogy
THE ANALOGY #Think of a wound. Gold is skin never cut. Aluminium is a cut that scabs over instantly and holds — knock the scab off and a new one forms before anything gets in. Iron is a cut whose scab is too thick to stay stuck, so it keeps lifting away and taking a layer with it, and the wound works steadily inward.
A scab is repair and rust is not — nothing is restored, only converted — and a wound heals to a stop, whereas iron in damp air has no endpoint short of the last atom of metal.
Clarifying the model
THE MODEL #Passivation — the self-sealing oxide — is not unique to aluminium. Stainless steel works the same way, using chromium to grow the film iron cannot grow for itself, which is also why stainless corrodes where that film is starved of the oxygen it needs to reform, as in a crevice or under a gasket.
The mechanism is also a tool. If corrosion is a circuit, you can choose where the electron-releasing end sits: bolt zinc to a steel hull, or coat the steel in it, and the zinc gives up its electrons preferentially, consumed so that the steel is spared. That is galvanising and sacrificial protection — a deliberate use of the same asymmetry that ruins the nail.
One honesty note: neat anode and cathode patches are a simplification. On a real surface those regions shift, swap, and overlap as the water film, oxygen supply, and rust layer change, and predicting corrosion rates in the field remains substantially empirical.
A picture of it
THE PICTURE #How to readStart at the rounded terminal and follow the two diamonds, the only real forks. The first asks whether the metal will react at all — gold exits here, which is the whole of its nobility. The second asks what the oxide does once formed, and matters more: aluminium exits to a sealed, starved surface, while iron falls through to the flaking branch. That branch loops back to the anode box on clean metal, and the loop is the point — it is why rusting never stops. The parallelogram is the circuit itself, and the cylinder hanging off it is salt, which does not start the reaction but makes the film conduct.
What became clearer
WHAT CLEARED #Corrosion is an electrical circuit running on one piece of metal, electrons crossing through the metal and ions through a water film. Gold escapes it by never entering the transaction. Aluminium enters eagerly and then chokes it off with an oxide that seals. Iron enters and produces an oxide too bulky to stay put — so the metal is never covered, and the circuit re-forms on whatever the flaking exposes.
Where to go next
ONWARD #- Why two different metals in contact corrode faster than either alone.
- Why crevices and waterline marks corrode first, when the metal there is identical.
Key terms
TERMS #| Term | What it means |
|---|---|
| Anode | the region where metal atoms give up electrons and dissolve as ions. |
| Cathode | the region where those electrons are consumed, in damp air by oxygen and water forming hydroxide. |
| Electrolyte | the conducting liquid film that lets ions complete the circuit, which salt makes far more effective. |
| Passivation | a thin, adherent oxide film that blocks further reaction, as on aluminium and stainless steel. |
| Sacrificial protection | coating with, or attaching, a metal that corrodes preferentially, such as zinc on steel. |
Every term the collection defines is gathered in the glossary.