A raw starch granule won't dissolve in cold water. Heat is what forces it open — water pushes into the structure, the granule swells to several times its size, and past a certain point it ruptures, spilling long chains of starch into the surrounding liquid. That's gelatinization, and it's the one mechanism behind a roux, a slurry, and a pudding alike. What catches people off guard is that the process runs backward once the pot cools. The same molecules that just leaked out start finding each other again, re-bonding into a tighter mesh that squeezes some of the water back out. That second act is called retrogradation, and it's why a sauce sets thicker in the fridge than it looked in the pan — and why a loaf of bread goes hard on the counter days after it left the oven.
What's actually inside a granule of starch
Raw starch — cornstarch, wheat flour, the starch inside a raw potato — is built from two shapes of the same glucose polymer, packed into a granule. Amylose is long and mostly straight, like a coil of uncooked spaghetti. Amylopectin is shorter and branches constantly, closer to a bramble than a strand. Inside the granule, both get wound into tight, ordered rings held together by hydrogen bonds, alternating with looser, disordered layers between them. That packing is dense enough to keep cold water out completely. Stir flour or cornstarch into a cold liquid and nothing happens — the granule sits there fully intact until it meets real heat.
Heat breaks the crystal, and the temperature that takes depends on the starch
Heat does two things to a granule at once. It shakes loose the hydrogen bonds holding the crystalline rings together, and it lets water push into the gaps that opens up. Past a specific temperature, the structure gives way fast enough that the granule can't hold its shape anymore — it swells, and the liquid around it thickens as a direct result. That temperature isn't the same for every starch. It depends on how tightly a given plant's crystalline regions are packed to begin with, and lab measurements of four common starches heated slowly in water found a real spread between them.
| Starch | Onset | Fully swollen (peak) |
|---|---|---|
| Wheat | 126°F (52°C) | 139°F (59°C) |
| Potato | 136°F (58°C) | 143°F (62°C) |
| Corn | 140°F (60°C) | 152°F (67°C) |
| Rice | 145°F (63°C) | 160°F (71°C) |
Wheat starch — already sitting in a spoonful of flour — starts swelling barely above a hard simmer. Rice starch needs a liquid close to a boil before it does the same thing, a full 21°F gap between the loosest starch of the four and the most tightly bound.
Bursting, not just swelling, is what actually thickens the liquid
A granule that's merely swollen adds some body on its own, but the real thickening happens when it bursts. Past its peak, the granule's structure gives out completely, and long amylose chains that were locked inside spill into the surrounding water, tangling with each other and with what's left of the ruptured granules. That tangled network is what a spoon feels as resistance — not water anymore, but water threaded through with molecules too large and too intertwined to move past each other freely.
Why the same sauce sets thicker once it's cold
Cooling doesn't undo any of this. It runs it in a different direction. Amylose molecules that were kept apart by heat and motion slow down as the liquid cools, and within hours they start finding each other again, re-forming some of the same hydrogen bonds that held the raw granule together in the first place. That re-bonding is retrogradation, and it squeezes free water out from between the chains as it tightens — the same water that turns up as a thin, weepy layer on top of cold pudding or a refrigerated pan sauce. A cornstarch- or flour-thickened sauce that looked perfectly loose in the pan can set into something closer to a gel by the time it's fully chilled, with no extra starch added at any point.
Waxy starches skip that second act almost entirely
Not every starch carries the same amount of amylose to begin with, and that difference decides how firmly a starch sets once it cools. Wheat, corn, and most rice varieties run somewhere around a quarter of their weight in amylose, which gives their gels enough straight-chain material to re-bond into a firm set. Waxy corn starch and glutinous rice are bred to have almost none of it — closer to pure amylopectin — and amylopectin's branching shape makes it far harder for chains to line back up into an ordered mesh. Tapioca sits in between: less amylose than wheat or corn, and what's there resists re-bonding, which is the practical reason a tapioca-thickened pie filling or a bubble tea pearl stays glossy and pliable instead of setting into a firm gel days later.
The slower version of this is why bread goes stale
Bread crumb is gelatinized starch, set by the oven's heat instead of a stovetop's. The moment a loaf starts to cool, the same retrogradation clock that firms up a refrigerated sauce starts running inside it — just far slower, because most of a wheat crumb is amylopectin, and amylopectin needs days rather than hours to recrystallize. As it does, the crumb's structure tightens and its texture goes from tender to firm, independent of how much water is actually still inside it. That's the part that surprises people: a stale loaf hasn't necessarily dried out. Sealing it in a bag slows moisture loss, not retrogradation, which is why bread wrapped airtight still goes hard on schedule.
The refrigerator makes bread stale faster, not slower
Retrogradation doesn't run at one speed. It runs fastest in a narrow band just above freezing, roughly the temperature of a home refrigerator, and slows down sharply both above room temperature and below the point where the crumb actually freezes solid. Put a loaf in the fridge to keep it fresh and it goes stale several times faster than the same loaf left on the counter, because refrigerator temperature sits close to the worst possible speed for the reaction. A true freezer works the opposite way: cold enough that water locks into ice and the starch molecules barely move, retrogradation almost stops entirely — the actual reason freezing is the one storage method that holds bread's texture for more than a few days.