Thermodynamics

The kernel that is a pressure cooker until it bursts

A popcorn kernel is a drop of water sealed in a nearly steam-tight shell. Heated, the water cannot boil off; it climbs water's boiling curve instead, and by the time the shell gives way, near 180 °C, the steam inside is at about ten bar — five times what a kitchen pressure cooker holds. When the shell fails the water is eighty degrees hotter than water can be at one atmosphere, and the stored heat turns half of it to steam in an instant. Almost all of that energy leaves as steam. The foam it leaves behind takes a few per cent; the leap across the pan, about a millijoule.

Assumes: A boiling point is a pressure, not a temperature · The part of the curve no fluid follows

Only a few kinds of maize pop. Field corn, sweet corn and the flint varieties grown for meal all scorch in a hot pan without much more than a crack; popcorn turns itself inside out into a white foam thirty or forty times its own volume. The difference is not the starch, which is much the same in all of them, nor the water, about a seventh of the kernel’s weight in each. It is the hull. A popcorn kernel’s outer layer, the pericarp, is unusually strong and close to impermeable to steam, and a kernel is popcorn because it is a pressure vessel.

Seen that way, popping is one of the cleanest demonstrations in the kitchen of the fact that a boiling point is a pressure, and of what superheated water does when the pressure holding it is suddenly removed.

The boiling curve, climbed in a sealed shell

Water in an open pan boils at 100 °C because that is the temperature at which its vapour pressure equals the atmosphere’s; heat it more and it does not get hotter, it turns to steam. Seal it in and there is nowhere for the steam to go. Each extra degree raises the vapour pressure along the saturation curve, which Clausius and Clapeyron showed must rise steeply because boiling absorbs so much latent heat for such a large change of volume, and the water stays liquid at temperatures it could never reach in an open pan.

A kernel is a pressure cooker with a ten-bar lid. The pressure at which water boils against its temperature, from 90 to 200 °C. Water sealed in a container cannot boil off; it heats, and its vapour pressure climbs this curve until the container fails. A domestic pressure cooker holds about 2.0 bar absolute and cooks at 121 °C. A popcorn kernel's hull holds about ten — 10.0 bar at 180 °C, the temperature at which kernels are measured to pop — and the water inside it is then 80 degrees hotter than water can be in an open pan.
Fig. 1 The pressure at which water boils against its temperature, from 90 to 200 °C. Water sealed in a container cannot boil off; it heats, and its vapour pressure climbs this curve until the container fails. A domestic pressure cooker holds about 2.0 bar absolute and cooks at 121 °C. A popcorn kernel’s hull holds about ten — 10.0 bar at 180 °C, the temperature at which kernels are measured to pop.

A kitchen pressure cooker exploits the curve modestly: its lid and valve hold about one bar above the atmosphere, and its water cooks at 121 °C, the temperature an autoclave uses to sterilise. A popcorn kernel goes much further. Emmanuel Virot and Alexandre Ponomarenko filmed kernels popping in an oven in 2015 and found they popped at about 180 °C, with little spread from kernel to kernel; at 180 °C water’s vapour pressure is ten bar. Earlier measurements of the pressure at bursting give similar values. The kernel’s hull is holding nine atmospheres of steam over the outside air when it fails — more than the pressure in a road-racing bicycle tyre, held by a shell a twentieth of a millimetre thick.

The water inside is not a pool. It is distributed through the kernel’s starch, partly bound to it, and its vapour pressure is somewhat below that of free water at the same temperature. Treating it as free water, as the curve does, puts an upper bound on the pressure at each temperature, and the agreement with the measured burst pressures says the bound is not far off by the time the kernel is near bursting.

Why the window is narrow

The boiling curve is steep at 180 °C, and its steepness has a consequence for how evenly a batch pops. Clausius and Clapeyron’s slope is the latent heat divided by the temperature and by the change in volume on boiling, and near 180 °C it is about a fifth of a bar per degree. A hull ten per cent stronger than its neighbour, holding nine-tenths of a bar more, therefore pops only about four degrees later. Large variations in hull strength from kernel to kernel are compressed by the curve into a narrow band of popping temperatures, which is part of why Virot and Ponomarenko found so little spread.

The latent heat in that slope is large for a reason the entropy every liquid gains on boiling set out: boiling multiplies a liquid’s volume by more than a thousand and so gains it an entropy of roughly ten gas constants per mole, whatever the liquid. Water gains more than most, because its molecules are held together by hydrogen bonds, and a steeper vapour-pressure curve is the result. A popcorn kernel filled with a liquid of smaller latent heat would show a gentler curve and a wider spread of popping temperatures for the same hulls.

The hull decides the temperature

A thin spherical shell of radius rr and wall thickness hh, holding a pressure difference ΔP\Delta P, carries a stress in its wall of

σ=ΔP r2h,\sigma = \frac{\Delta P\, r}{2h},

the hoop stress of a pressure vessel, and it bursts when that stress reaches the wall’s strength. The cell that bursts by geometry met the same formula in a red blood cell swelling in fresh water, where the membrane’s tension, not its thickness, is what has a limit. For a kernel two and a half millimetres in radius with a hull fifty micrometres thick, nine bar of steam puts 22.5 megapascals of stress in the wall — about the tensile strength of a fibrous plant tissue — and that value is used here as the hull’s strength.

Halve the hull and the kernel pops twenty-five degrees cooler. The temperature at which a model kernel bursts, against the thickness of its hull, for a 2.5 mm kernel whose hull fails at a wall stress of 22.5 MPa. The pressure a thin shell can hold is proportional to its thickness, and through the boiling curve each pressure is a temperature: 156 °C for a 25 µm hull, 180 °C for 50 µm, 210 °C for 100 µm. At altitude the outside pressure is lower and the kernel bursts slightly cooler — by about a degree — because what the hull holds is the difference. The strength here is a model value chosen to put a 50 µm hull at 180 °C.
Fig. 2 The temperature at which a model kernel bursts, against the thickness of its hull, for a 2.5 mm kernel whose hull fails at a wall stress of 22.5 MPa. The pressure a thin shell can hold is proportional to its thickness, and through the boiling curve each pressure is a temperature: 156 °C for a 25 µm hull, 180 °C for 50 µm, 210 °C for 100 µm. At altitude the kernel bursts slightly cooler — by about a degree.

The pressure a hull can hold is proportional to its thickness, and the boiling curve turns each pressure into a temperature. A hull half as thick bursts at 156 °C, one twice as thick at 210 °C. The variation is gentler than the pressure’s, because the curve steepens as it rises, but it is real, and it explains a good deal about how a batch pops. Kernels differ in their hull thickness and in small flaws, and so pop at different times as the pan heats. Kernels whose hulls are cracked never build pressure at all: their steam leaks away as fast as it forms, and they are the “old maids” left in the bottom of the bowl. Kernels stored too long dry out, and a dry kernel’s water is held more tightly by the starch, so its vapour pressure lags the curve and it reaches the hull’s limit later, if at all — which is why popcorn is sold in sealed containers and pops best at about fourteen per cent moisture.

The figure for altitude is a small surprise. The hull holds the difference between the inside and outside pressure, so at 1,600 metres, where the air is at 0.83 bar, the steam inside needs to reach only 0.83 plus nine bar to burst it, and the kernel pops about a degree cooler. What changes more at altitude is what happens next.

The flash

When the hull fails, the pressure on the water inside falls to the atmosphere’s in a fraction of a millisecond. The water is at 180 °C and the atmosphere says it may be at 100 °C at most. It is superheated — a liquid on the part of the curve no fluid follows at equilibrium — and it cannot stay there: it boils explosively throughout its volume until the kernel has cooled to 100 °C. The heat that does the boiling is the heat the kernel stored above 100 °C, in the water and in the starch around it, and it is spent as latent heat.

The heat stored above boiling turns half the water to steam in an instant. The share of a kernel's water that turns to steam the moment its hull fails, against the temperature at which it fails, for a kernel of 14% moisture at sea level. Above 100 °C the kernel stores sensible heat that the sealed water could not spend on boiling; at the burst the pressure drops to one atmosphere and that heat evaporates water until the kernel is back at 100 °C. At 180 °C it is 50% of the water, about 118 cm³ of steam per gram of kernel — more than the 30–45 cm³ a gram of good popcorn puffs into, so most of the steam escapes through the foam it has just made.
Fig. 3 The share of a kernel’s water that turns to steam the moment its hull fails, against the temperature at which it fails, for a kernel of 14% moisture at sea level. At 180 °C it is 50% of the water, about 118 cm³ of steam per gram of kernel — more than the 30–45 cm³ a gram of good popcorn puffs into.

For a kernel of fourteen per cent moisture bursting at 180 °C, the stored heat evaporates about half its water. The starch, which is most of the kernel’s mass, holds most of that heat, so the flash is driven largely by the kernel’s dry matter giving up its heat to the water inside it. Half the water, a fourteenth of a gram of steam per gram of kernel, occupies 118 cubic centimetres at 100 °C and one atmosphere. The starch, softened by heat and water into a molten gel, is blown into a foam by this expansion, and the foam sets as it cools and dries: a gram of good popcorn ends up occupying thirty to forty-five cubic centimetres. The steam is three times the flake, and most of it escapes, which is why popcorn steams as it pops.

Because the starch holds most of the stored heat, the amount of steam the flash makes hardly depends on how much water the kernel had. A kernel of ten per cent moisture bursting at 180 °C flashes 0.066 grams of steam per gram; one of eighteen per cent, 0.073 — the drier kernel flashes a larger share of a smaller amount of water, and the totals nearly agree. What the moisture does change is whether the kernel reaches the burst at all, and in what state: too little water, held tightly by the starch, and the vapour pressure lags the curve so that the hull is never stressed enough before the starch scorches; too much, and the starch is too soft and wet when it is blown out, and the foam collapses before it sets. The optimum near fourteen per cent that growers and packers aim for is a compromise between those two failures, not a maximum of the flash.

This is the same event that drives a geyser. The geyser that remembers one eruption followed water superheated at the bottom of a conduit by the weight of the water above it; when a little is thrown out, the pressure drops, the superheated water flashes, and the eruption feeds itself. The kernel’s lid is its hull instead of a column of water, and its eruption takes a millisecond instead of minutes, but the physics is the same, and so is the conclusion: the violence is set by how far above the open boiling point the water was held.

Superheat held by a lid, not by a lack of bubbles

Water can be superheated in a second way, and the contrast shows what the hull is doing. Heated in a smooth, clean cup in a microwave oven, water can rise several degrees above 100 °C at atmospheric pressure without boiling, because a steam bubble has to begin somewhere, and the barrier a new phase has to climb — the surface energy of a tiny bubble — keeps it from beginning in the body of a clean liquid. Jog the cup or drop in a spoon and the water erupts. That superheat is metastable: it is held by the difficulty of starting.

The kernel’s superheat is not like that. Before the burst its water is not superheated at all; it is at its boiling point for the pressure the hull is holding, sitting exactly on the curve. The superheat appears only at the burst, when the pressure falls by a factor of ten in a fraction of a millisecond, and then nothing holds it: the starch granules and their gelatinised surroundings are full of the irregularities a bubble needs to start on. The kernel goes from the stable side of the curve to far beyond it, and boils at once, everywhere. The difference is between a liquid that has not yet found a way to boil and one whose permission to stay liquid has just been withdrawn.

Lower the pressure outside and the same kernel makes more steam

The flash depends on the outside pressure in two ways that add. A lower outside pressure lowers the temperature the kernel can fall to, so more of its stored heat is available to evaporate water; and steam at a lower pressure fills more space per gram.

Lower the pressure outside and the same kernel makes more steam. The volume of steam a kernel flashes per gram when its hull fails, against the pressure of the air it bursts into. A lower outside pressure lowers the boiling point the kernel falls to, so more of its stored heat goes into evaporation, and each gram of steam fills more space: 118 cm³ at sea level, 150 at 1,600 m and 477 at 0.3 bar. Popping under partial vacuum is reported to give larger flakes, which this direction agrees with; how much of the steam's extra volume the starch foam actually keeps is not something this balance can say.
Fig. 4 The volume of steam a kernel flashes per gram when its hull fails, against the pressure of the air it bursts into. A lower outside pressure lowers the boiling point the kernel falls to, so more of its stored heat goes into evaporation, and each gram of steam fills more space: 118 cm³ at sea level, 150 at 1,600 m and 477 at 0.3 bar.

At 1,600 metres the steam per gram rises by about a quarter. Popped into a partial vacuum of 0.3 bar, the same kernel would make four times the steam volume, and popping under reduced pressure is reported to give larger flakes. The balance here says only that the driving gas is larger; how much of it the starch foam can keep before its bubbles rupture and the steam escapes is a matter of the foam’s strength while it sets, which the balance knows nothing about.

Where the energy goes

The flash’s energy can be followed to the end, and the accounting is lopsided.

Almost all the energy of the pop leaves as steam. Where the energy goes when a gram of popcorn pops, on a logarithmic axis. Heating the kernel to 180 °C takes about 314 J. At the burst, 157 J of that is spent turning water to steam. Making room for the 40 cm³ the flake grows to, against the atmosphere, takes only 4.1 J — 3% of the flash — and throwing the popped kernel 5 cm into the air takes about half a millijoule. The pop is an explosion of steam that happens to leave a foam behind it.
Fig. 5 Where the energy goes when a gram of popcorn pops, on a logarithmic axis. Heating the kernel to 180 °C takes about 314 J. At the burst, 157 J of that is spent turning water to steam. Making room for the 40 cm³ the flake grows to, against the atmosphere, takes only 4.1 J — 3% of the flash — and throwing the popped kernel 5 cm into the air takes about half a millijoule.

Heating a gram of kernel from room temperature to 180 °C takes about 314 joules. When the hull fails, 157 joules of that is spent turning water to steam. Pushing back the atmosphere to make room for forty cubic centimetres of flake takes four joules, three per cent of the flash. And the kernel’s famous leap — Virot and Ponomarenko found that a starchy leg pushed out from the burst point kicks the kernel off the pan, and that the pop’s sound comes from steam released at the same moment, not from the hull cracking — takes about half a millijoule for a five-centimetre jump. The pop is an explosion of steam, and the flake is what the explosion leaves behind.

The same imbalance explains why popcorn is not a good engine. The water was heated to 180 °C at a pressure of ten bar, and an engine taking heat at that temperature and rejecting it at 100 °C could in principle convert about eighteen per cent of the heat into work, by the ceiling on every engine. The kernel converts three per cent of the flash into inflating its foam, because the expansion happens through a burst rather than against a piston, and almost all the steam’s capacity for work is thrown away as it rushes out.

Field corn, flint corn and the microwave

The comparison with ordinary maize is the experiment that shows the hull is the point. A field-corn kernel has as much water and as much starch, and in a hot pan it gets just as hot, but its pericarp is thinner and more permeable, and it loses its steam through the hull as fast as the steam forms. It never builds pressure, never superheats when it opens, and merely scorches. Popcorn varieties also have a larger proportion of hard, glassy endosperm, the vitreous starch whose tightly packed granules hold the water in place and let the gel stretch into a foam without tearing at once — the softer, floury starch of field corn makes a poor foam even when a kernel is made to burst. But the hull comes first: a popcorn kernel with a pinhole drilled in it behaves like field corn.

A microwave oven pops corn differently and to the same end. It heats the water inside the kernel directly, rather than heating the kernel from outside, so the inside reaches the burst temperature without the hull having been baked first — one reason microwave popcorn bags carry a thin metallised patch that absorbs microwaves and heats the oil and the kernels in contact with it from outside as well. Microwaves heat a dish unevenly, and the oven that cooks in spots followed the standing-wave pattern that makes them do so; a bag of kernels spread across the floor of the oven averages over it, and the turntable averages over it again.

What the vessel model leaves out

Treating the kernel as a sphere of free water inside a uniform shell is a strong simplification. The water is dispersed through starch granules, its vapour pressure lowered by binding, and some of it migrates and escapes before the burst; the hull is not a uniform shell but thicker in places and weaker at the tip where the kernel was attached to the cob, which is where kernels usually fail. The hull strength used here is a model value chosen to put a fifty-micrometre hull at 180 °C, not a measured property. The flash calculation assumes all the kernel’s heat above 100 °C is available to the water in the instant of the burst, which overstates the flash for the outer parts of the kernel, which lose heat to the pan and air as they open. And the heat to warm the kernel ignores the water that evaporates through the hull during heating, a few per cent of the kernel’s mass, which is part of why over-dried popcorn pops poorly.

Still open: what decides the shape of the flake

Popcorn flakes come in two broad shapes, the irregular “butterfly” with wings and the rounder “mushroom”, and they behave differently — mushrooms survive coating with caramel, butterflies have more surface for butter. Which shape a kernel makes depends on the variety, the moisture, and the heating rate, but the mechanics of how the starch foam expands through the burst opening, sets, and fractures, and why it settles into one shape or the other, is understood only qualitatively. It is a problem in the rheology of a gelatinised starch foam expanding against its own setting, at a timescale of milliseconds, and it has had less attention than the economic value of the answer would suggest.

A popcorn kernel’s hull seals in its water, which cannot boil and instead climbs water’s boiling curve — 10 bar at the 180 °C at which kernels pop, five times a pressure cooker — until the hull’s hoop stress reaches its strength, so the hull’s thickness sets the popping temperature, 156 °C for half the thickness and 210 °C for double. When it bursts, the kernel’s heat above 100 °C flashes half its water to steam, 118 cm³ a gram, three times the flake it leaves; inflating the foam takes 3% of that energy and the leap half a millijoule, and the rest leaves as steam — the same flash that drives a geyser, with a twentieth of a millimetre of hull in place of the water column.

Part 19 of 19

This essay is one argument about Phase change. The others:

The objects named here

The third axis, after the field and the reading path: the things themselves, and every essay that touches each one.

Clausius clapeyronFlash evaporationHoop stressLatent heatSuperheatingVapour pressure