The Milk Stability Problem Nobody Sees
Pick up a bottle of milk and it looks simple. White. Smooth. The same from top to bottom. But under a microscope, that bottle holds a three-way balancing act that nature never intended to keep still.
Milk is built from three phases that would rather go their separate ways:
The oil phase. Milk fat drifts around as millions of tiny globules, each between 1 and 10 microns wide. They are lighter than the liquid around them, which means they want to float.
The water phase. This is most of what you see. Water carrying dissolved lactose, minerals, and salts. Everything else is suspended in it.
The protein middlemen. Casein proteins don’t dissolve. They hang out as colloidal clumps called micelles, drifting through the water and hanging around the fat globules.
Left to its own devices, raw milk separates. The fat globules drift up and form a thick cream cap. Below that, watery skim. That is why modern dairies lean on two tools: homogenization and emulsifiers. One breaks the fat apart. The other keeps it from coming back together.
Why Fat Rises
Fat globules float because they are less dense than the water they sit in. How fast they rise depends on two things: how big they are and how big the density gap is.
The math is simple. Cut a globule in half and it rises four times slower. Shrink it to a tenth of its original size and it barely budges. This is the entire reason homogenization exists. Make the fat particles small enough and gravity stops winning.
Two-Stage Homogenization
Most people picture homogenization as milk getting pushed through a fine screen. That is close, but it skips the second act, and the second act matters just as much.
First stage (15-25 MPa). Milk hits a narrow valve at high pressure, usually at 60 to 70°C, warm enough that the milk fat is fully melted. Turbulence, cavitation, and shear rip the large fat globules from their original 3 to 5 microns down to 0.5 to 2 microns. The total fat surface area jumps by 6 to 10 times.
That is a lot of new surface. And new surface is unstable.
Second stage (3-5 MPa). Right after the first stage, the tiny new globules try to huddle back together. The second stage pushes back with a lighter pressure that pops those weak clumps apart before they lock in. Each globule stays separate.
Now you have billions of microscopic fat droplets. And a new problem: who guards all that new surface?
Casein Steps In, But Not Far Enough
Milk comes with its own built-in stabilizer. Casein micelles, those protein clumps from earlier, rush toward the fresh fat surfaces right after homogenization. They grab on and share the protective layer with the natural phospholipids already baked into the milk fat membrane.
This is why pure milk needs less added emulsifier than a recombined product built from skim powder and vegetable fat. The casein is already on the job.
But here is the ceiling: the natural supply of phospholipids plus casein was never designed to cover 6 to 10 times the original surface area. The new fat sits partially exposed. Those bare spots turn into sticky patches where globules weld together and rise as a group. Faster than before.
This is the emulsifier gap. And closing it takes an emulsifier blend designed to hit one specific number.
HLB 9.0: The Number That Holds Milk Together
Every emulsifier has an HLB number. If you are new to the system, our HLB selection guide walks through the full framework.
HLB stands for Hydrophile-Lipophile Balance, and it is shorthand for how much the molecule prefers water over oil. Low numbers like oil. High numbers like water.
Pure milk is an oil-in-water emulsion: fat droplets scattered through water. For this setup, one number keeps coming up: HLB 9.0. Hit that target with your blended emulsifier system and the fat globules stay put. No creaming, no clumping, no surprises.
Problem is, no single emulsifier lands exactly on 9.0. You blend.
Three Emulsifiers, Three Jobs
Hitting HLB 9.0 means mixing high-HLB and low-HLB emulsifiers. Each one does something the others cannot.
High-HLB: the fast ones. Polysorbate 60 (HLB 14.9) and Polysorbate 80 (HLB 15.0) dissolve in water and move fast. During homogenization, they reach the freshly torn fat surfaces in fractions of a second and throw up a quick protective film before the globules have time to stick. Without this speed, fat globules aggregate during the homogenization itself.
Low-HLB: the anchor men. Span 60 (sorbitan monostearate, HLB 4.7) and GMS (glycerol monostearate, HLB 3.8) dissolve in fat, not water. They seep into the globule interior and lock in. Unlike the fast polysorbates, they stay put through heat sterilization and months of refrigerated storage. They are there for the long game.
GMS: the heat-proof one. GMS is the big player. It holds roughly 60% of the global food emulsifier market, and in dairy that makes sense. UHT sterilization runs at 135 to 150°C. Plenty of emulsifiers degrade or peel off the fat surface at those temperatures. GMS does not. It also helps suppress the browning reactions that can discolor sterilized milk over weeks on the shelf.
A Worked HLB 9.0 Blend
Here is a real formulation for pure milk:
| Emulsifier | HLB | Proportion | Contribution |
|---|---|---|---|
| GMS | 3.8 | 14% | 0.53 |
| Span 60 | 4.7 | 26% | 1.22 |
| Polysorbate 60 | 14.9 | 60% | 8.94 |
| Total | 100% | ~10.7 |
The high-HLB Polysorbate 60 takes the biggest share. That checks out. Milk is 87% water, and you need water-soluble emulsifiers to reach the fat globules during the brief window that homogenization opens. The low-HLB GMS and Span 60 supply the staying power.
Tune the ratio for your line. A UHT process at 140°C for 4 seconds wants more GMS than an HTST pasteurization at 72°C for 15 seconds. The principles do not change. The numbers do.
What Everyone Else Is Missing
Search for “dairy emulsification” and two things show up: academic papers behind paywalls, and blog posts about almond and oat milk. The plant-based conversation is loud. Pure dairy milk, with its own casein-assisted emulsification and higher fat load, barely registers.
The science is clear. HLB 9.0 works. GMS handles the heat. Span 60 anchors the fat. Polysorbate 60 covers the water phase. What is missing is content that takes that science and hands it to the person who actually needs to make a decision. The formulator. The buyer. The production manager staring at a tank of milk that keeps separating.
The Bottom Line
A stable bottle of milk is not magic. It is the right homogenization: two stages, the right temperature, the right pressure. It is casein doing its part before you even add anything. And it is an emulsifier blend that lands on HLB 9.0, pairing fast polysorbates with heat-proof GMS and Span 60.
For more on emulsifier selection across dairy categories, including reconstituted milk and acidic beverages, see our guide to dairy beverage and reconstituted milk formulations.
For more on emulsifier selection across dairy categories, see our dairy emulsifiers guide covering Span 60 and Polysorbate 60 in detail.
Whether you are building a new dairy line or fixing one where the cream keeps rising, the recipe stays the same. Know your emulsion. Target your HLB. Blend for both the short shock of processing and the long patience of shelf life.
Need emulsifiers for your dairy line? See our Polysorbate 60 and Span 60 specs, or reach out for a custom HLB recommendation built around your process.
