Cast-iron seasoning

Why does a cast-iron pan become non-stick only after years of cooking, and lose it in a single wash?

The folk account of a seasoned pan is that the oil has soaked into the metal, the way water soaks into a sponge, and that years of cooking drive it deeper. It is a comfortable picture and it explains the slow improvement. But iron is not porous in any way that oil could occupy, and if the surface were merely oily it would wipe off on the first cloth. Something is bonded to that pan that behaves like neither iron nor oil. What is it?

Begin with what oil is. A cooking oil is largely triglycerides — fatty acid chains, and the useful ones for our purposes are unsaturated, meaning their carbon chains contain double bonds. A double bond is a reactive site: it holds the neighbouring carbon-hydrogen bonds more weakly than a saturated chain does, and oxygen can get at them.

Now heat a very thin film of such an oil in air. Oxygen attacks those weak positions and forms unstable peroxides, which break apart into radicals — fragments with an unpaired electron, desperate to bond with something. What they find is other fatty acid chains, and they join to them. Each joining creates a new radical further along, which finds another chain, and so on.

Follow that a few million times and ask what you have. Not a thicker oil. Chains linked to chains in every direction form a continuous three-dimensional network — a solid polymer, hard, insoluble, no longer a liquid in any sense. This is not special to cooking; it is exactly how linseed oil turns into the tough film of an oil painting or a varnish, linseed being simply a very unsaturated oil.

So a seasoned surface is a plastic, and that explains the puzzle we started with. Nothing soaked in; a layer was built on top, bonded to the metal surface, and it is neither iron nor oil but a third material with its own properties — one of which is a low-energy surface that food proteins bond to poorly. That is where the non-stick behaviour comes from.

Why should it take years, then? Because a thick coat of oil polymerises badly: the outside cures into a skin while the inside stays soft, and the result flakes. Only a film wiped down to almost nothing cures throughout. So each cooking session that leaves a trace of fat and then heats it adds one very thin, sound layer — and what accumulates is not depth of oil but number of good layers, each filling the microscopic roughness of the cast surface a little more. Slow accumulation is not an inefficiency of the process; it is the only way the process works.

Which oil, then? The instinct is to reach for smoke point, since recipes talk about heating past it. But smoke point measures when an oil starts to break down and volatilise, and that is not the reaction we want. What we want is double bonds to oxidise and cross-link, so the property that predicts a good film is the degree of unsaturation — what the iodine value of an oil measures. Highly unsaturated oils such as flaxseed cure fastest, though the film is notably brittle and prone to flaking off in sheets. Saturated fats, having few double bonds, polymerise poorly however hot you get them.

The second half of the original question needs correcting rather than answering, because its premise is largely wrong. A single wash does not remove seasoning. The advice never to use soap on cast iron is real advice, but it dates from an era when household soap was made with lye and carried free alkali, and strong alkali does attack the ester linkages in a polymerised film. Modern dish detergents are neutral synthetic surfactants; they lift grease, which is the loose fat sitting on top, and leave the cured polymer alone. A cured film is not soluble in anything a sink contains.

So what does destroy it? Strong alkali still does — oven cleaner and lye baths are the standard way to strip a pan deliberately, which is itself the proof of the mechanism. Steel wool removes it mechanically. Long simmering of acidic foods such as tomato or wine eats through it and then attacks the iron beneath. And enough heat, such as an oven's self-cleaning cycle, burns it off entirely. What people experience as "lost it in one wash" is almost always one of these, most often a scouring pad or a soak that let water sit on bare patches and rust them.

Two honest limits. The polymerisation chemistry is well established, but the precise nature of the bond between film and metal is less settled than the cross-linking itself, and likely involves the oxide layer rather than the iron directly. And "non-stick" here is relative: a well-seasoned pan is far better than bare iron and not the equal of a manufactured fluoropolymer coating.

• Why flaxseed seasoning cures beautifully and then flakes, and what brittleness in a cross-linked film depends on.
• Why a pan left wet rusts through the seasoning from underneath.

Every term the collection defines is gathered in the glossary.