Sandiness from Lactose — Supersaturation and Crystallization

Why does the tongue meet sand when the MSNF — the milk solids-not-fat — is pushed too high? The culprit is a crystal of lactose. This piece follows the other crystallization that runs in the unfrozen phase, the counterpart to the recrystallization of ice (Approfondimento 3).

Advanced Chapter 3 said the ceiling on MSNF is set by the solubility of lactose, and warned that too much brings the defect of sandiness. Advanced Chapter 13 put it on the list of things to diagnose. This piece digs into what that sand really is — how lactose reaches supersaturation, and what kind of crystal it becomes. Like the recrystallization of ice in Approfondimento 3, this too is a story of crystallization running in the unfrozen phase.

Spot illustration of A crystal of α-lactose. The crystal behind the grittiness.

Lactose dissolves markedly poorly

Lactose is a disaccharide of dextrose and galactose joined together, and it is found in the milk of animals and nowhere else. Among the sugars it has one peculiar property: a markedly low solubility.

How low is that? In 100 grams of water at 20 °C, some 200 grams of sucrose will dissolve. Against that, what first dissolves of the α form of lactose is only about 8 grams. One twenty-fifth, in fact.

There is another step to the story. The α form, once dissolved, crosses over in water to the β form — mutarotation. The β form dissolves more readily than the α, so as the α is used up, room opens for more α to dissolve. By the time equilibrium is reached, the effective solubility runs higher than that first 8 grams. Even so it stays around a tenth of sucrose. The working figure of Advanced Chapter 3 — one part lactose to six parts water at 15 °C — is this value after equilibrium. By either measure, lactose dissolves worse than the other sugars by an order of magnitude. That is where this piece starts.

Recall that Advanced Chapter 3 said lactose always comes along with the MSNF, and the danger shows itself. The more MSNF is added to strengthen the frame, the more this reluctant sugar piles up in the unfrozen phase.

Fig. L3-6-1　How well they dissolve at 20 °C, per 100 grams of water. Measured by what first dissolves of the α form, the gap runs to twenty-five times. Lactose reaches saturation on a slight rise.

Freeze concentration makes the supersaturation

What settles the problem is freezing. As Approfondimento 1 showed, when water leaves as ice, what remains in the unfrozen phase grows stronger — freeze concentration. Lactose is carried along in it. A sugar of low solubility to begin with, robbed of water and made stronger still, passes its solubility easily. This is supersaturation.

Supersaturation is a strained state: more is dissolved than can stay dissolved. Thermodynamically it is unstable, and the excess lactose will sooner or later come out as crystal. With the MSNF in proportion, the unfrozen phase never reaches saturation and no crystal appears. Push past the ceiling, though, and the freeze-concentrated unfrozen phase reaches supersaturation, and the trigger for crystallization is pulled. Advanced Chapter 3 puts a ceiling on MSNF, and Advanced Chapter 5 finds that ceiling by calculation, to keep this supersaturation away.

Fig. L3-6-2　Freezing drives lactose into supersaturation. The more water the ice takes, the stronger the lactose in the unfrozen phase grows, and past the line of solubility the crystals come out. The more MSNF, the sooner the line is crossed.

The sand is α-lactose monohydrate

So what is the crystal that comes out? Lactose crystallizes from solution in two forms, split at 93.5 °C. Below that line it is always the α form as a monohydrate — α-lactose monohydrate. A freezer is of course in that range, so the lactose crystal that grows in gelato is, without exception, this one.

What makes the crystal troublesome is that it is very hard. Unlike sucrose, which dissolves and vanishes in the mouth, a crystal of α-lactose monohydrate is hard, angular, and slow to redissolve. Its shape is distinctive too — a pointed, arrowhead shape, known at a glance under the microscope. Hard, pointed grains scattered over the tongue cannot go unfelt. This is the feel of sand, the substance of the defect. In sensory work as well, sandiness names, plainly, the perception of lactose crystals.

It resembles the way coarse ice crystals make things gritty (Approfondimento 3), but this is a crystal of sugar, governed by another measure — solubility. And unlike ice, it will not melt at the temperature of the mouth. The grittiness of ice goes away if it is held on the tongue; the sand of lactose stays to the end.

Here the mutarotation of the opening tells again. Dissolved lactose stands in equilibrium between the α and β forms, but what leaves as crystal is only the α form, the less soluble one. As the α goes, the equilibrium is broken, and β crosses over to α to fill the gap. Then more leaves. On dissolving, mutarotation worked to raise the amount that dissolves. On coming out, it works to keep the crystal growing. One mechanism, acting in opposite directions on the way out and on the way back.

It bites on a delay

The defect of sandiness has a delay of its own. Even once supersaturation is reached, the crystal takes time to grow to a size the tongue can feel. So sandiness usually shows itself not on the day of making but after some days of storage. The column in Advanced Chapter 3 put it as the effect on the day of making, the side effect some days later. It points at this lag in crystal growth.

And exactly as with the recrystallization of ice, swings in temperature speed this growth up. Every time the product temperature moves, lactose dissolves and comes out again, and the crystal grows larger. Sandiness too is a defect weak to heat shock (Approfondimento 3).

Never let the supersaturation happen

The remedy comes down to never pulling that trigger. And this defect has an advantage no other has — it can be prevented on paper, before it happens.

The first road is to calculate the lactose concentration of the unfrozen phase from the formulation. The lactose in the dairy ingredients is known. The factor of freeze concentration can be estimated from the ice fraction at the serving temperature (Approfondimento 1). So how strong the lactose in the unfrozen phase will get can be worked out before anything is mixed. If it comes out too high, lower the skim milk powder. About half of skim milk powder is lactose, so this is the rein that tells hardest. The ceiling formula of Advanced Chapter 5 is this calculation folded into a working procedure.

The second road, for when the solids have to go higher regardless, is the lactose-hydrolyzed product. Convert the lactose enzymatically to dextrose and galactose, and these dissolve far more readily than lactose. They never reach supersaturation, so sandiness is avoided (Advanced Chapter 3). But splitting them doubles the number of molecules. The freezing point and the sweetness both move, and that must not be forgotten (Advanced Chapter 2, Approfondimento 1).

There are also the auxiliary moves shared with Approfondimento 3 and 4. Raise the viscosity of the unfrozen phase with a stabilizer, and the diffusion of lactose molecules slows and crystal growth dulls. Hold the cold chain low and steady, and the growth itself is held back. Ice and lactose — the two crystallizations that run in the unfrozen phase are reined by the same two reins. One is the formulation, meaning the amount of lactose. The other is the process, meaning the steadiness of the temperature.