Reducing Sugars in Gelato: Why Your Pale Base Browns
Table of contents
A fiordilatte that leaves the pasteuriser the colour of weak tea was not burnt. It browned, and browning needs two partners: a protein and a sugar with a free carbonyl group. Milk always brings the protein. Which sugar you chose decides whether it brings the other half.
What Makes a Sugar Reducing
A reducing sugar is one whose ring can open to expose a free aldehyde or ketone group. That open form is chemically reactive, and it is what the classic Fehling and Benedict tests detect when they reduce copper in solution.
Dextrose is reducing. Fructose is reducing. Lactose and maltose are reducing, because in both disaccharides only one of the two anomeric carbons is locked into the glycosidic bond and the other stays free.
Sucrose is the exception that matters most in a gelato lab. Its glucose and fructose halves are joined anomeric carbon to anomeric carbon, so neither ring can open. Sucrose is non reducing and, on its own, contributes nothing to this reaction.
That single fact explains a great deal of bench experience. Two bases with identical sugar percentages can leave the pasteuriser two different colours if one is built on sucrose and the other on invert sugar, which is nothing but glucose and fructose set free.
| Sugar | Reducing | Browning contribution |
|---|---|---|
| Sucrose | No | None on its own |
| Maltodextrin, DE under 20 | Very few free ends | Negligible |
| Glucose syrup DE 38 | Yes | Moderate, rises with DE |
| Dextrose | Yes | High |
| Fructose | Yes | High, the fastest of the common sugars |
| Invert sugar and honey | Yes, both halves | High |
| Lactose, from milk solids | Yes | Moderate, but always present |
Fructose sits at the top of that list for a structural reason. It exists in solution with a larger share of its molecules in the open chain reactive form than glucose does, so gram for gram it reacts faster. Any sweetener built on it, from honey to invert sugar to high fructose syrups, carries the same tendency.
The Reaction Itself
The carbonyl group of a reducing sugar condenses with a free amino group, most often the epsilon amino group of lysine in milk protein. The product rearranges into an Amadori compound, and from there a cascade of dehydrations and fragmentations ends in brown nitrogenous polymers called melanoidins.
This is the Maillard reaction, and it is entirely distinct from caramelisation. Caramelisation is sugar decomposing on its own and needs high temperatures, around 160 °C for sucrose. Maillard runs at pasteurisation temperatures and, slowly, at refrigeration temperatures too. Marshall, Goff and Hartel treat it in Ice Cream as the standard explanation for cooked flavour and colour development in heat treated mixes.
Four variables set the rate. Temperature and time raise it together, which is why holding is more damaging than peak temperature. Higher pH accelerates it, since the amino group must be unprotonated to react. Intermediate water activity favours it, which is why a dried milk powder in a warm store browns faster than the liquid milk it came from. And concentration matters: more reducing sugar, more protein, more reaction.
Quick reference. Sucrose is non reducing and does not brown by itself. Dextrose, fructose, invert sugar, honey and lactose all do. Milk protein supplies the amino partner in every dairy base, so the sugar choice is the only half of the pair you control.
Where the Reducing Sugars Hide
The reducing sugar in most pale bases was never added deliberately. It came in with the milk solids.
Lactose is roughly 52 g per 100 g of non fat dry milk according to USDA FoodData Central, against about 36 g of protein in the same 100 g. A base carrying skim milk powder to reach its MSNF target therefore holds both halves of the reaction in high concentration, at the exact ratio that favours it.
The powder itself carries a history. Low heat skim milk powder is dried under conditions that leave whey proteins largely undenatured and Maillard products low. Medium and high heat powders have already run part of the reaction in the drying tower, and a base built on them starts browner and browns further. Whey protein concentrate is more reactive still, because whey proteins carry a high proportion of available lysine.
Then there are the sugars deliberately chosen for their PAC. Honey is largely fructose and glucose, both reducing. Invert sugar and high DE glucose syrup are rich in free glucose. Dextrose equivalent is literally a measure of reducing sugar content, expressed as a percentage of dextrose on dry basis, so DE reads directly as a browning index: DE 60 browns; maltodextrin below DE 20 barely does.
One more source appears in fruit work. Sucrose hydrolyses to glucose and fructose under heat and acid, so a sucrose only sorbetto base held hot at low pH creates its own reducing sugars while it waits.
When Browning Is a Defect and When It Is the Product
The reaction is not an enemy. It is the whole point of toasted milk, of dulce de leche, of the roasted depth in hazelnut and of the colour of a caramel base. Those products are built by pushing exactly the variables listed above.
It becomes a defect in three places. Pale flavours lose their identity: fiordilatte, ricotta, lemon and pistachio all read as stale or cooked once the colour drifts. Delicate aromatics are masked by the caramel and biscuit notes that come with the pigment. And the reaction consumes lysine, which is a real if minor nutritional cost in a high MSNF product.
The line between the two cases is intent. If the browning was not designed, it is telling you the mix spent longer hot than it needed to.
Controlling It
Heat load is the first lever, and the cheapest. A short high temperature hold does less browning damage than a long low temperature one at equivalent lethality, which is one practical argument for the equipment discussed in the pasteurisation deep dive. Cool to ageing temperature quickly rather than letting the vat coast down, and age at 4 °C.
Sugar choice is the second. If a pale base browns and the recipe leans on invert sugar or a high DE syrup for its PAC, move part of that load onto sucrose and recover the missing PAC with dextrose only to the extent the colour tolerates. Reformulating toward maltodextrin for solids, rather than toward more milk powder, cuts both the lactose and the reducing sugar in one move.
MSNF is the third. Every extra point of milk powder adds lactose and lysine together. If the base is already near the sandiness ceiling, the same point of solids taken as maltodextrin browns less and crystallises less.
Storage is the last and the most often ignored. Powders held warm brown in the bag, and a mix aged overnight at 8 °C browns measurably more than the same mix aged at 4 °C. Both changes cost nothing.
Related Concepts
- Maillard reaction in gelato for the flavour side of the same chemistry
- Browning in gelato for the troubleshooting checklist
- Lactose in gelato for the sugar that arrives uninvited
- Dextrose equivalent explained for reading DE as a browning index
- Skim milk powder for heat classes and how to choose one
- Invert sugar for the highest risk sweetener on the shelf
