Starch vs Gum in Gelato: Which One Stabilises Better
Table of contents
Two shops make gelato with identical milk, sugar and fat. One thickens with cornstarch, the other with a pinch of locust bean gum. Both are stabilising the same mix, but they are solving different halves of the problem, and only one of them survives a bad freezer.
Twenty-five grams on the left does roughly what two grams does on the right.
They are not competing for the same job
Both starch and gum are polysaccharides, and both end up described as stabilisers on a spec sheet. That is where the similarity stops.
Starch works by bulk. Heat it in water and the granule gelatinises, swelling to many times its volume and physically soaking up free water. The viscosity you feel comes from a suspension of swollen granules crowding each other. To get a measurable effect you need serious weight: 20 to 30 g per kilogram of mix in a Sicilian-style crema, as covered in cornstarch in gelato.
A gum works by chemistry, in solution. Galactomannans like locust bean gum and guar dissolve into long hydrated chains that tangle with each other and raise the viscosity of the water phase enormously at tiny concentrations. Two grams per kilogram is a normal dose. Goff and Hartel put the whole stabiliser and emulsifier blend of a commercial ice cream mix in the 0.2 to 0.5 percent band, and that figure is for the blend, not for one gum.
Ten times the dose for the same apparent thickness. That difference is not an inefficiency; it is the whole story.
Where the ice actually forms
A frozen gelato is not solid. At a serving temperature near minus 12 °C, a meaningful share of the water is still liquid, held as a concentrated syrup of sugars, salts and proteins between the ice crystals. That unfrozen serum is where the damage happens.
Every time the tub warms slightly, a little ice melts back into the serum. Every time it cools again, that water refreezes, and it refreezes preferentially onto the crystals that already exist rather than forming new ones. Crystals grow at the expense of their neighbours. Enough cycles and the texture goes from smooth to coarse. That is ice recrystallisation, and it is what heat shock does to a tub in a display case.
A gum lives in that serum phase and makes it viscous, which slows the diffusion of water molecules towards the growing crystals. It does not stop recrystallisation; it buys time, and time is what a display case is short of.
Starch is not really in that phase. Its swollen granules are a dispersed solid, binding water inside themselves rather than thickening the syrup around the ice. It contributes body and it lowers the amount of free water available in the first place, but it is not standing between the water and the crystals in the same way.
| Starch | Gum | |
|---|---|---|
| Typical dose per kg | 20 to 30 g | 1.5 to 4 g |
| Adds total solids | Yes, ~88% dry matter | Negligible |
| Adds PAC | No | No |
| Main mechanism | Swollen granules bind free water | Serum viscosity slows water diffusion |
| Resists heat shock | Weakly | Strongly |
| Failure mode if overdosed | Pasty, floury, gluey | Gummy, rubbery, slow-melting |
| Needs cooking | Yes, above ~63 °C | Depends on the gum |
| Label reads | Corn starch | Locust bean gum, guar gum |
Quick reference. Starch buys body and solids. Gum buys shelf life. If the tub is eaten the day it is churned, starch alone is fine; if it sits in a case for a week, it is not.
Figure 1 — The gum works in the syrup between the crystals; the starch works inside its own granules.
What each one does to the balance sheet
Starch is a solids ingredient and has to be entered as one. Commercial maize starch arrives at roughly 10 to 12 percent moisture, so 25 g of it lands about 22 g of dry matter in the mix. In a base already sitting near the top of its total solids window, that is enough to tip it into heaviness on its own.
Gum is not. At 2 g per kilogram the solids contribution is a rounding error, which is exactly why gums are the tool of choice in a low-fat or low-solids base that needs body it cannot afford to buy with dry matter.
Neither moves the freezing point. Freezing point depression counts dissolved particles, and both of these are huge molecules present in tiny molar amounts, so PAC stays where your sugars put it. Anyone who tells you a stabiliser softens the scoop is describing the texture, not the thermodynamics.
Hydration is where gums go wrong
Starch is forgiving in one direction and unforgiving in the other: it does nothing until it is properly cooked, and once cooked it stays cooked. Gums are the opposite. Each one hydrates on its own schedule, and a gum that has not hydrated is money thrown away.
Guar and xanthan hydrate in cold water and will thicken a mix straight from the blender. Locust bean gum needs to be taken above roughly 80 °C to develop fully, which is why it belongs in a high pasteurisation cycle and disappoints anyone who stirs it into a cold base. Tara gum sits between the two. The stabiliser hydration temperature guide has the whole table.
The other classic gum failure is dispersion. Dropped in as a lump, a galactomannan wets on the outside and forms a gel skin around dry powder that never dissolves. Pre-blend it with sugar at twenty parts sugar to one part gum, or shear it in with a stick blender.
The clean-label argument, honestly
The most common reason artisans reach for starch is not texture. It is the ingredient list. "Corn starch" reads like food; "locust bean gum" reads like a factory, even though it is milled carob seed and has been in Mediterranean kitchens far longer than most of what sits next to it.
That is a real commercial consideration and it deserves a straight answer rather than a lecture. If your customers read labels and starch keeps them comfortable, use starch, hold the tubs to a short rotation, and accept that the shelf life is shorter. The Sicilian tradition ran that way for a century, backed by a fast turnover that never gave heat shock a chance.
If tubs sit for days, the honest options are a small gum dose, a lot more egg yolk, or a customer who eventually finds the gelato coarse. Pick deliberately.
Use both, in most cases
The good answer is usually not one or the other. A base carrying 15 to 20 g of starch per kilogram for body plus 1 to 1.5 g of a galactomannan for heat-shock protection gets the mouthfeel of the starch route and most of the durability of the gum route, at a dose low enough that nothing reads as gummy.
Overdo either and you can taste it. Too much starch and the gelato turns floury and dulls its own flavour release. Too much gum and it goes elastic, melts strangely on the tongue and earns the complaint covered in why gelato turns gummy. Start low, and change one variable per batch.
Related Concepts
- Do I need stabilisers? for whether either is necessary at all
- Stabiliser blend recipe for building your own gum mix
- Tapioca starch as the cleanest-flavoured starch option
- Xanthan gum for the cold-hydrating alternative
