Why dairy beverages separate, and what to check first

A dairy beverage that separates on shelf is one of the most common stability complaints in the category, and one of the most commonly misdiagnosed. The instinct is to increase the stabiliser. Sometimes that is right. Often the separation has a cause that more stabiliser will not reach.

The first useful step is simply to look at where the separation sits.

A layer on top: fat creaming

Milk is an oil-in-water emulsion, with fat globules dispersed through the continuous phase. The University of Guelph's dairy science text describes what happens without intervention: left to stand, the fat rises and forms a cream layer.

Homogenisation is the standard defence. Passing milk under high pressure through a small orifice reduces the average diameter of the fat globules and increases their number and surface area. The same source attributes the improved stability of homogenised milk to three factors: a smaller mean globule diameter, which is a term in Stokes' law, a narrower size distribution so that most globules rise at a similar speed, and an increase in globule density.

Stokes' law is the useful mental model here. The speed at which a globule rises depends on its size, on the density difference between the globule and the liquid around it, and on the viscosity of that liquid. Every fix for creaming works by changing one of those three.

What to check

• Whether the product was homogenised under the conditions intended, and whether the equipment is performing as it should
• Whether any fat was added after the homogenisation step
• Whether the continuous phase has the viscosity the formulation assumes

A layer at the bottom: particles settling

Sediment is the same physics in reverse. Particles denser than the liquid around them, such as cocoa, settle downwards. In chocolate and cocoa milk this is the defining stability problem.

A 2025 study in the Journal of Food Science examined exactly this, identifying the stability failures of milk-based cocoa beverages as sedimentation of cocoa particles, formation of lumps, and light and dark coloured layers. It describes the stabilising mechanism of kappa-carrageenan as a specific electrostatic interaction between the sulfate groups of carrageenan and casein micelles, which builds a network capable of holding cocoa particles in suspension.

Within that study, kappa-carrageenan was tested between 0.01% and 0.04%, and sedimentation reached zero once the concentration exceeded 0.025%. Homogenisation pressure, varied between 5 and 34 MPa, had a statistically insignificant effect on sedimentation.

That last finding is also a useful warning. If cocoa is settling, raising homogenisation pressure may do little, because the problem is suspension rather than particle size in the emulsion. It is a good example of why diagnosis comes before adjustment.

What to check

• What the sediment actually is: cocoa, a mineral fortificant, fibre, or protein that has come out of solution
• Whether the stabiliser was fully dispersed and hydrated before heat treatment
• Whether a raw material lot changed around the time the problem started

Bands through the bottle

Light and dark layers indicate that the system is partially structured but not uniformly so. This pattern often sits between the two failures above, and it is worth treating as its own diagnosis rather than as a milder version of either.

A diagnostic order that saves time

1. Identify which pattern you have, and photograph it at a fixed time after production.
2. Note when it first appears. Separation within days and separation after several weeks usually have different causes.
3. Check the process record for the affected batches before touching the formulation.
4. Establish whether anything changed: a raw material lot, a supplier, a line, a heat treatment setting.
5. Change one variable at a time, against a control produced the same day.

Following that order is slower than raising the stabiliser dose on the next batch. It is also far more likely to fix the problem permanently.