A raw egg white looks like nothing much: a clear, slippery gel. It is about 90% water. Almost all of the rest is protein, and each protein is a long chain of amino acids folded up into a tight little ball. The balls drift around in the water without sticking to each other, and they are too small to scatter light. That is why raw white is see-through.
Heat changes that. As the water warms, the molecules shake harder, until the weak bonds holding a ball in shape give way and the chain flops open. An unfolded chain has sticky stretches that used to be tucked inside. They grab other open chains, and soon there is a three-dimensional net running through the whole white, with the water trapped in its holes. The net is solid, and its clumps are big enough to scatter light. The clear liquid has become an opaque white solid.
Cooks call this setting. Chemists call it coagulation. Unlike melting ice, it does not undo itself when the egg cools.
Here is the part that makes eggs tricky. The white is not one protein but a mixture, and the yolk is a different mixture again. Each one unfolds at its own temperature.
| Around | What gives way | What you see |
|---|---|---|
| 62 °C | Ovotransferrin, about 12% of the white's protein | The white turns milky and just holds together |
| 65–70 °C | The yolk's proteins | Yolk thickens to a jam, then sets |
| 80 °C | Ovalbumin, about 54% of the white's protein | The white turns firm |
Read that table again and something odd stands out. The yolk sets at a lower temperature than most of the white. So how does a soft-boiled egg manage a firm white around a runny yolk? Try it.
Tap or drag on the egg to move the probe. Lime dots are ovotransferrin, pale dots are ovalbumin; the white's other proteins are left out. The boiling mode treats a fridge-cold egg as a simple ball of water, so the times are a rough guide, not a recipe.
In boiling mode, drag the timer and watch the probe readings. Water at 100 °C is far hotter than anything in the egg needs, but heat can only creep inward through the white. After a few minutes the outer white is well past 80 °C and firmly set, while the centre of the yolk has barely warmed up. A soft-boiled egg is not an egg cooked to one temperature. It is an egg caught halfway through heating, with a steep temperature slope from shell to centre.
That is why the timer matters so much, and why a minute either way changes breakfast. Time does not really cook the egg. Temperature does. Time only decides how far the heat has travelled.
Now switch to the gentle bath and set it to about 66 °C. In this mode the whole egg sits at one temperature for long enough to even out. The ovotransferrin has let go, so the white is milky and fragile. The ovalbumin has not, so the white never firms up. The yolk, meanwhile, is inside its own setting range and turns thick and custardy.
The result is a boiled egg turned inside out: a soft, barely-there white around a yolk that is the firmer of the two. This is the Japanese onsen tamago, the "hot spring egg", traditionally made by lowering eggs into naturally hot spring water. Restaurants now do the same thing with a temperature-controlled water bath, where a single degree makes a visible difference to the yolk.
Setting is not the end of the story. Keep heating a set white and the protein net keeps tightening, with more and more bonds pulling the chains together. A tight net squeezes out the water it was holding. The white turns from tender to rubbery, and the yolk from moist to dry and crumbly.
Long cooking also frees sulfur from the white's proteins. It meets iron from the yolk right where the two touch and forms ferrous sulfide, a grey-green compound. That is the ring around the yolk of an overdone hard-boiled egg. It is harmless, just unlovely. The two fixes are the ones cooks already know: do not overcook, and cool the eggs quickly in cold water once they are done.
An egg is a stack of proteins that each give way at a different temperature, packed inside a shape that heats unevenly. Every style of egg, from onsen to soft-boiled to hard, is one answer to the same question: which proteins have unfolded, and where?