Dynamic vs Static Freezing: How Gelato Ice Crystals Form
Table of contents
Every gelato is frozen twice. First dynamically, in a freezer that scrapes and stirs, then statically, in a cabinet or blast freezer where nothing moves. The two stages make ice in completely different ways, and the texture you serve is decided by how much of the work each one does.
Two Ways to Turn Water Into Ice
Dynamic freezing happens while the mix is agitated. In a batch freezer or a continuous scraped surface heat exchanger, a dasher sweeps blades across a refrigerated wall, lifts off the ice that forms there and throws it back into the moving mix. Air is whipped in at the same time.
Static freezing, which food scientists call quiescent freezing, happens with no agitation at all. Water loses heat through the container walls and ice grows wherever it can. Hardening a tray of freshly drawn gelato is static freezing. So is an ice lolly, a no-churn tub in a home freezer, or a granita between rakings.
Neither is "better". A professional process uses both, in a fixed order: dynamic first to create the crystal population, static second to finish the job without destroying it.
What Happens Inside the Batch Freezer
The best single summary of the mechanism is Cook and Hartel's review, Mechanisms of Ice Crystallization in Ice Cream Production (Comprehensive Reviews in Food Science and Food Safety, 2010). Its central point is that nucleation and growth happen in different places.
- At the wall. Only the thin layer of mix touching the refrigerated barrel is cold enough to form new crystals. In industrial continuous freezers the wall sits around −26 to −28 °C with refrigerant near −30 °C, so supercooling there is large and ice nucleates immediately.
- The scrape. The dasher blade removes that frozen layer and disperses it into the bulk of the freezer, which is much warmer, close to the mix's own freezing point.
- In the bulk. Here some crystals melt and others grow, a process the review calls ripening. The branched, dendritic shapes formed at the wall round off.
- At the draw. The mix leaves as a population of small, rounded, disc-shaped crystals.
The numbers are what make dynamic freezing so valuable. Surveying the published studies, Cook and Hartel report an average crystal size of 20 to 30 µm at the freezer exit, with about 50% of the water frozen. Marshall, Goff and Hartel's Ice Cream places the draw temperature for ice cream at roughly −5 to −6 °C; our draw temperature guide covers how artisan gelato machines compare.
Friction is a hidden cost. The dasher puts heat back into the product: measurements cited in the review found that 15% to 40% of the heat removed was frictional heat from the dasher, depending on dasher speed and draw temperature. That is one reason a longer churn is not a better churn.
What Happens When Nothing Moves
Take the agitation away and three things change at once.
- Fewer nuclei. Without a scraper constantly harvesting fresh crystals from the cold surface, nucleation happens once, mostly at the container walls, and then growth dominates.
- Bigger, branched crystals. The review notes that dendrites may persist intact for a long time in quiescent systems. In one experiment it cites, a 40% sucrose solution frozen without stirring formed hexagonal crystals of 100 to 450 µm before they turned dendritic.
- No air. Nothing whips the mix, so a still-frozen base has essentially no overrun unless the air was folded in beforehand, as in a semifreddo.
Compare those sizes with the sensory threshold. The same review states that crystals larger than about 50 µm give a grainy texture. A still-frozen gelato base overshoots that line, which is why no-churn methods rely on heavy fat, sugar or whipped air to disguise the ice, or break it up mechanically afterwards, as the food processor method and the Pacojet do.
Hardening: Static Freezing Done Right
After the draw, roughly half the water is still liquid. It freezes during hardening, and it cannot form new crystals there: it deposits onto the ones already present.
Quick reference. Dynamic freezing sets the crystal count: about 50% of the water frozen and crystals of 20 to 30 µm at the draw. Static hardening adds mass to those crystals, growing them 30% to 40% to about 25 to 45 µm, with about 75% of the water frozen. Still freezing from liquid gives crystals well above the 50 µm grainy threshold.
Cook and Hartel give the figures: during hardening, crystals typically grow by 30% to 40% to an average of about 25 to 45 µm, and in the hardened product about 75% of the water is frozen. The logic that follows is simple and explains most practical advice about freezing: the more crystals there are at the draw, the more ice mass is shared out among them during hardening, and the smaller each one stays.
That is why the review concludes that a lower draw temperature leads to smaller crystals: less water is left to freeze statically. It is also why speed matters in the cabinet. Higher temperatures accelerate recrystallization, so quick hardening limits growth. A blast chiller and shallow pans both shorten the time the product spends in that vulnerable zone.
Dynamic vs Static at a Glance
| Dynamic freezing | Static (quiescent) freezing | |
|---|---|---|
| Where it happens | Batch or continuous scraped surface freezer | Hardening cabinet, blast freezer, home freezer, mould |
| Nucleation | Continuous, at the cold wall | Mostly once, at container walls |
| Crystal shape | Small rounded discs | Dendrites and large plates if starting from liquid |
| Typical size | 20 to 30 µm at the draw | Grows existing crystals; from liquid, well over 50 µm |
| Air | Incorporated, overrun built | None added |
| Water frozen | About 50% at the draw | Up to about 75% in hardened product |
| Main risk | Long residence time, frictional heat | Slow hardening, heat shock |
What This Means at the Machine
The research carries some counterintuitive lessons.
Time in the barrel matters more than wall temperature. Russell and colleagues (1999), and later Drewett and Hartel (2007), both found that residence time had the biggest effect on final crystal size. Shorter residence time gave smaller crystals. In a batch machine, that means loading the right quantity for the barrel and not letting the finished gelato keep turning after it is ready, which only feeds ripening and warms it through friction. The batch vs continuous freezer comparison shows how the two designs handle this.
Formulation decides how much water is left to freeze. Sugars depress the freezing point, so at any given temperature a better balanced mix has less ice. That reduces the load on static hardening. Freezing point depression explains the mechanism, and the PAC figure in your recipe is how you control it.
Static freezing never stops being a risk. Every temperature swing in the display case partly melts small crystals and refreezes that water onto large ones. That is ice recrystallization, and repeated swings are what we call heat shock. Stabilizers slow it; they do not stop it.
When Static Freezing Is the Goal
Coarse ice is not always a defect. A Sicilian granita is frozen almost statically and raked at intervals so that large crystals form and stay separate. Granita and sorbet share a recipe logic but sit at opposite ends of this spectrum. Semifreddo gets its lightness from air whipped in before freezing, then sets statically in the mould, as covered in semifreddo vs gelato.
For gelato, though, the rule holds: build as many small crystals as possible while the mix is moving, then freeze the rest as fast and as steadily as you can once it stops.
Balance before you freeze. Build the recipe in the Free Gelato Balancing App so the mix reaches the machine with the right sugars and solids. The less free water left at the draw, the less your hardening stage has to get right.
Related Concepts
- Ice crystal size and gelato texture, the 50 µm threshold and how to stay under it.
- Draw temperature in gelato, where dynamic freezing ends.
- Mantecazione, the Italian batch freezing step in practice.
- Why is my gelato icy?, diagnosing coarse texture.
