The Salt and Ice Bath Method for Making Gelato at Home

Before mechanical refrigeration, every gelato was frozen in a bucket of ice and salt. The method still works, and on a home bench it beats every other machine-free approach, because it is the only one that supplies a heat sink colder than the mix while you agitate. Here is the physics and the procedure.

The whole apparatus: an inner bowl, an outer bucket, crushed ice, coarse salt and a spatula.

Why An Ice Bath Freezes Anything At All

Plain ice sits at 0 °C. A gelato base does not freeze at 0 °C: dissolved sugars and salts depress its freezing point to somewhere around -2.5 to -3.5 °C for a balanced formula, which is the whole subject of freezing point depression and the reason PAC exists as a number. Pack a bowl of mix in plain ice and nothing happens. The bath is warmer than the target.

Salt fixes that. Sprinkled onto ice, it dissolves into the thin film of liquid water always present at the surface, and the resulting brine has a freezing point below 0 °C. The ice is now above its own melting point relative to that brine, so it melts. Melting costs energy: 334 kJ for every kilogram of ice converted to water, and that energy has to come from somewhere. It comes from the surroundings, which means the brine, the bucket, and the bowl of mix sitting in the middle of it.

The bath therefore drives itself downward as long as ice remains and salt remains undissolved. This is not the salt "making cold". It is ice melting at a temperature far below where it would normally melt, and stealing the latent heat to do it.

The Eutectic Floor And What It Means For You

The descent is not unlimited. As the brine concentrates, its freezing point falls along a curve until brine and ice reach equilibrium together. For sodium chloride and water that point is 23.3% salt by mass at -21.1 °C, the eutectic of the system. No mixture of ice and table salt, at any ratio, in any bucket, will go below that.

Quick reference. The NaCl and water eutectic is -21.1 °C at 23.3% salt by mass. That is the absolute floor of a salt-and-ice bath; in a real bucket, expect -12 to -18 °C, which is already colder than a domestic freezer.

Figure 1 — Freezing point of a sodium chloride solution against salt concentration, ending at the eutectic.

Read that table against your mix and the working window becomes obvious. You need the bath meaningfully colder than the base's own freezing point, or heat transfer is too slow and the mix simply thickens without freezing. You also do not want the bath at its absolute floor for the whole run: an aggressively cold wall freezes a thick shell on the inside of the bowl faster than you can scrape it off, and the result is the coarse, shattery texture covered in why gelato turns icy.

The target is a bath somewhere around -12 to -18 °C. That gives a driving force of roughly 10 to 15 degrees against the mix, which is the same order of magnitude a domestic batch freezer works with.

Building The Bath

Salt. Coarse rock salt or coarse sea salt, not fine table salt. Coarse crystals dissolve gradually, which keeps brine forming steadily through the run instead of dumping the whole depression in the first two minutes and then fading. Iodised table salt works chemically but exhausts itself fast. This is a different job from salt in the mix itself, where a few tenths of a percent are there for flavour.

Ice. Crushed, not cubed. Contact area is everything: crushed ice packs against the bowl wall and gives the brine somewhere to sit, while cubes leave insulating air gaps and freeze unevenly.

Ratio. Start at roughly one part salt to six parts crushed ice by weight, layered rather than dumped: ice, scatter of salt, ice, scatter of salt. For a 1 litre mix you will want about 3 kg of ice and 500 g of salt, plus a reserve of both, because you will top up.

Vessels. A stainless steel inner bowl, as thin-walled as you have, sitting inside a larger insulated bucket or a second bowl wrapped in a towel. Stainless conducts; plastic and ceramic do not, and a ceramic bowl will roughly double your churn time.

Thermometer. Non-negotiable. Put a probe in the bath, not in the mix. Everything above is a starting point, and the bath temperature is the only thing that tells you whether the starting point was right for your ice, your salt and your room.

The Churn

Chill the mix to refrigerator temperature first and, if the recipe calls for it, let it age. Cold mix into a cold bath is the whole game; warm mix into a cold bath wastes half your ice on the first few degrees. The ageing and maturation step is worth keeping even without a machine.

Pour the mix in no more than 4 cm deep. Depth is the enemy: the bowl only cools from its wall, and a deep pool has a warm core you cannot reach.

Then agitate continuously. A flexible spatula, working the frozen film off the wall and folding it back into the centre, for the full run. This is not stirring for the sake of it. Agitation does the two things mantecazione does in a machine: it keeps ice crystals small by breaking them up as they nucleate, and it beats in the air that becomes overrun. Stop for five minutes and you get a solid rim and a liquid middle.

Expect 20 to 40 minutes for a litre. The mix goes from liquid to the consistency of thick cream at about the 10 minute mark, then to soft-serve, and you draw it there. Chasing scoopable texture in the bucket overworks it. Transfer to a shallow container, cover the surface, and firm it in the freezer for an hour or two, then serve near the right serving temperature.

Top up salt and ice roughly halfway through. The bath warms as ice is consumed, and a bath that drifts back above -8 °C will stall the whole batch.

Where This Method Beats The Alternatives

Four machine-free routes are in common use, and we compare them in detail in no-churn methods compared. Ranked by the texture they can actually reach:

The ice bath wins for a structural reason rather than a sentimental one. It is the only home method where a cold wall and continuous agitation happen at the same time. Freezer air is a poor conductor, so the scrape-and-beat route spends most of its cycle with no agitation at all; the food processor method has excellent agitation but the cold is all pre-loaded and running out; the condensed-milk route never freezes dynamically at all. Simultaneity is what keeps ice crystals small, and small crystals are the entire definition of creamy rather than icy.

The cost is labour. Forty minutes of continuous spatula work for a litre is real effort, and it is the honest reason the method fell out of use rather than any deficiency in the result.

Three Failures And Their Causes

The mix stays liquid after 30 minutes. The bath is not cold enough. Check the probe: if it reads above -8 °C, you are short of salt, short of ice, or both, and the ratio needs to move toward one part salt to four or five parts ice. A close second cause is a ceramic or plastic inner bowl.

A thick frozen shell forms and the centre stays soft. You are not scraping often enough, or the mix is too deep. Work the wall constantly and keep the pool under 4 cm.

The result is dense and heavy. Too little air was incorporated, usually from folding too gently or from drawing too late. Fold vigorously through the early thickening phase, when the mix is still fluid enough to trap air, and draw at soft-serve rather than waiting for scoopable. A standard white base hand-churned this way lands well below machine overrun, and that is expected.

Related Concepts

• No-Churn Gelato Methods Compared puts this method against the other three.
• Freezing Point Depression in Gelato explains why the mix resists freezing.
• How to Make Gelato Without a Machine is the wider overview.
• Ice Crystal Size and Gelato Texture covers what agitation is protecting.
• What Temperature to Serve Gelato closes the loop after hardening.