Shear-thickening food pastes
A Socratic walk-through of shear-thickening food pastes — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a cornflour paste flow like cream until you stir it hard, then set solid under the spoon?
Mix cornflour into water until it pours thickly off a spoon. Tip the bowl and it runs. Now stab the spoon in fast and it refuses — not resists, refuses, as though you had struck a floor. Pull the spoon out slowly and it comes free, and the surface heals behind it as if nothing had happened.
The refusal is the puzzle. Nothing was added and nothing set. The paste has two mechanical characters and switches between them in milliseconds, in both directions. What is different about the fast case that is not different about the slow one?
Reasoning it through
REASONING #Begin with what is in the bowl: starch granules, packed very densely, with just enough water to fill the gaps between them. Around six parts solid to four parts water by volume is the interesting range; drop much below that and you get a paste that is merely thick, and the effect disappears.
Why should density matter so much? Because at that packing there is barely any room. For grains to move past each other, the pile must locally expand — a shearing bed of grains has to open up to let neighbours slide by. Reynolds called this dilatancy in the 1880s: press a footprint into wet beach sand and the surface goes dry and pale, because the sand has expanded and swallowed the water that was on top.
Hold that thought and ask what stops the granules from touching. At rest and under gentle shear, each pair is separated by a thin film of water, and pushing them together squeezes that film out — which takes time, because the water has to flow sideways through a narrowing gap. Slowly, there is time; the films survive, particles glide, the paste flows.
Now push hard. The force pressing pairs together rises with the stress you apply, and past some level it wins: the film ruptures and the granules make direct contact. And a contact is not merely an obstacle — it is a frictional one. Two touching grains can carry a sideways load; two lubricated grains cannot.
That single change is enough to explain the floor. How densely you can pack grains before they lock depends on whether they are slippery or grippy: frictionless spheres jam only near a volume fraction of about 0.64, frictional ones jam far lower, nearer 0.55. So a paste mixed at, say, 0.58 solids sits below the frictionless jamming point and above the frictional one. Stress it hard enough to convert lubricated contacts into frictional ones and you have moved the jamming threshold across the composition, rather than moving the composition. The suspension has not changed; the point at which it locks has.
What locks, mechanically? Chains of touching granules spanning from your spoon to the wall of the bowl, carrying the load as a temporary skeleton. Stop pushing and there is nothing holding the chains together — no bonds, only contact under load — so they collapse, water flows back into the gaps, and the paste is liquid again. That reversibility is the tell. A gel that set would stay set.
The analogy
THE ANALOGY #Think of a crowd leaving a stadium through a wide gate. Walking, people flow: each finds a gap, adjusts, keeps moving, and the exit rate rises with how briskly everyone walks. Now let the crowd surge. Bodies come into contact, and contact between people is frictional — shoulders lock, arches of pressure form across the gateway, and the crowd stops dead in an arch that no individual is strong enough to break. Ease back and the arch dissolves; nobody is stuck to anybody.
people brace deliberately and can hold an arch after the push stops, whereas starch granules have no such capacity — the structure exists only while the stress that made it is applied, which is why the effect is instantaneous in both directions.
Clarifying the model
THE MODEL #The habitual explanation, that stirring "makes the particles pile up into clusters", is not quite the mechanism. Hydrodynamic clusters are real and account for the mild thickening seen in ordinary thick suspensions, but the abrupt, orders-of-magnitude version is now attributed to that lubricated-to-frictional transition, worked out theoretically around 2014 and matched by experiments since. The difference is testable, and this is the falsification the account stands on: coat the particles so they slide more easily against one another, and at unchanged volume fraction the thickening should be pushed to higher stress or removed altogether. Add a lubricant and see the floor disappear, and the story survives. If altered surface friction made no difference while only particle concentration mattered, the frictional-contact account would be wrong and the crowding picture right.
A second, sharper test concerns which variable is in control. The paste is usually described as responding to speed, but the mechanism says the threshold is a stress. Shear the same paste in a thin gap and a wide one: at matched stress it should thicken in both, at matched rate it should not. That is the practical rule too — a slow spoon pressed hard can jam a paste that a fast spoon swept lightly does not.
Two honest limits. This is not thixotropy, which is time-dependent thinning of a structured fluid, nor is it any kind of setting: nothing is created and no chemistry occurs. And the numbers above are approximate on purpose. The critical volume fraction depends on particle shape, size spread, surface roughness and swelling, and starch granules swell in water, so a cornflour paste's true solid fraction drifts as it hydrates — which is why kitchen recipes for it are given by feel rather than by ratio, and why published thresholds differ between systems more than the effect itself does.
A picture of it
THE PICTURE #How to readStart at the top and read downward as force increases. The paste sits in the lubricated state whenever it is handled gently, and the branch to the right of it is the ordinary case of pouring. The move out of that state is triggered by stress, not by speed, and it happens in two steps worth separating: contacts turn frictional first, then those contacts link into chains that span from spoon to bowl and carry the load. The arrow back up is the whole point — the only thing holding the jammed state together is the stress itself, so removing it returns the paste to where it began, with no memory of the excursion.
What became clearer
WHAT CLEARED #Nothing in the bowl changes when the paste turns solid. What changes is the packing density at which the granules would lock, and stress changes it by breaking the water films that were keeping them slippery. Frictional grains jam at a looser packing than lubricated ones, and the paste is deliberately mixed to sit between the two — so it is liquid or solid depending on which kind of contact you have just forced it into. Stop forcing, and it forgets.
Where to go next
ONWARD #- Why chocolate, ketchup and cream behave in the opposite way, thinning as they are worked.
- How shear thickening limits the pumping and mixing rates achievable in a food plant, and what formulators change to suppress it.
Key terms
TERMS #| Term | What it means |
|---|---|
| Volume fraction | the share of a suspension's volume occupied by solid particles; the single most important variable for this behaviour. |
| Dilatancy | the requirement that a densely packed granular bed expand in order to shear, first described by Reynolds in wet sand. |
| Discontinuous shear thickening | an abrupt rise in viscosity, often by orders of magnitude, above a critical applied stress. |
| Jamming | the arrest of flow when particles form load-bearing contacts spanning the system; the packing at which it occurs is lower for frictional particles than frictionless ones. |
Every term the collection defines is gathered in the glossary.