Yes, carrageenan can be replaced in plant milk — but almost never one-for-one. Carrageenan does two jobs at once: it suspends particles and it adds body. Gellan gum replaces the first job at a lower dose (0.015–0.04% against 0.02–0.05%), and it does not replace the second at all. If you swap the carrageenan out and change nothing else, the drink will stay homogeneous and taste thinner. Budget for both changes: low acyl gellan at 0.02–0.03% plus 0.05–0.10% of a body-building hydrocolloid or extra solids. On material the swap is close to neutral: the gum itself drops by about a third while total stabiliser powder rises, and the two roughly cancel for most contract prices.
The short version. Gellan is a suspension system, not a thickener. Use it to keep cocoa, calcium and pulp in place through UHT and shelf life. Rebuild the mouthfeel separately, and reformulate the calcium and sequestrant at the same time — the gel network in a plant milk is controlled by free calcium, not by the gum alone.
Why the market is removing it
Three pressures, and they arrive together. Retail customers in the EU and North America have started writing "no carrageenan" into specifications, not because of a ruling against E407 in general but because the ingredient has a public-relations problem that brand owners would rather not defend. Infant formula regulation in the EU already restricts carrageenan, and buyers generalise that decision to adjacent categories. And clean-label programmes treat it as a synthetic-sounding additive even though it is extracted from red seaweed.
For a plant milk this lands on the formulator as a short brief: keep the calcium in suspension, keep the fat from creaming, keep the shelf life, lose the carrageenan, and do not let the product get thinner. That last clause is the one that breaks most first trials.
What carrageenan is actually doing in the pack
Carrageenan in a plant milk is usually a kappa/iota blend at 0.02–0.05%. It is doing four things at once, and only one of them is replaceable by a simple substitution:
- Suspension. Holding cocoa, calcium carbonate, vitamin premix and fruit pulp against gravity for 9–12 months. Kappa carrageenan forms a weak, thixotropic network with the calcium already present in the base.
- Body. A light viscosity and a rounded, slightly fatty mouthfeel. This is what consumers read as "milk-like" rather than "watery".
- Fat stability. Slowing creaming and limiting the size of the fat layer at the top of the pack.
- Protein protection. In neutral-pH systems, limiting aggregation of heat-denatured protein during UHT.
Gellan gum handles the first and partly the fourth, because low acyl gellan builds a calcium-crosslinked network that is far more efficient per gram than carrageenan's. It does essentially nothing for the second. A low acyl gellan solution at 0.03% has a viscosity close to water — the network is a gel at rest and a fluid under shear, which is exactly what you want for suspension and exactly the opposite of what you want for body. The same logic applies to the other carrageenan jobs, which is what the stabiliser system map sets out function by function.
The substitution table
Figures below are working ranges for a neutral plant milk at pH 6.4–7.2, UHT treated, ambient shelf life. Compare the carrageenan column with the two gellan columns: low acyl alone, and low acyl with a body builder.
| Parameter | Carrageenan (kappa/iota) | Gellan LA, suspension only | Gellan LA + body builder |
|---|---|---|---|
| Dose | 0.020–0.050% | 0.015–0.030% | 0.015–0.030% gellan + 0.05–0.10% acacia, CMC or citrus fibre |
| Hydration temperature | 75–85 °C | 85–95 °C | 85–95 °C (hydrate the gellan first, add the body builder after) |
| Set temperature | 40–55 °C, weak gel | 30–50 °C, ion dependent | 30–50 °C, ion dependent |
| Free calcium needed | Low — uses native calcium | Yes: 0.02–0.05% CaCl₂ if the base is soft water | Yes, same |
| Sequestrant | None normally | 0.03–0.06% sodium citrate to hold the set until filling | Same, and check the body builder for calcium binding |
| Working pH | 5.0–8.0, degrades below 4.3 | 4.0–8.0; gels hardest at pH 3–4 | Same |
| Viscosity contribution | Moderate | Near zero | Moderate to high, depending on the builder |
| Stabiliser powder per tonne | 0.35 kg | 0.22 kg gellan + 0.70 kg builder | Less gum, more total powder |
Any price stated here would be wrong for your contract, so compare the powder instead. Gellan delivers more suspension per gram of gum than carrageenan does, and it carries none of carrageenan's body — which means the gum goes down and the total powder goes up. Whether that nets out cheaper depends on two numbers only you have.
The dose arithmetic, without the prices
Take a 1,000-litre batch of oat base at 1.02 kg/L. Below is every figure that is a property of the formulation rather than a property of your purchasing:
| Line | Carrageenan route | Gellan route | Difference |
|---|---|---|---|
| Stabiliser dose | 0.035% | 0.022% gellan + 0.07% acacia | — |
| Gum per tonne | 0.35 kg | 0.22 kg | −37% |
| Total stabiliser powder per tonne | 0.35 kg | 0.92 kg | +163% |
| Break-even band | At a gellan price around two to three times the carrageenan price, the two effects roughly cancel | ||
| Label claim value | — | "Carrageenan free" | often the whole business case |
Two conclusions follow from the weights. The gum itself drops by roughly a third, which matters because gellan is priced above carrageenan per kilo. The total powder rises, because something has to carry the body that carrageenan was carrying on its own. Inside the two-to-three-times band those roughly cancel; above it the gum line rises and the case for switching has to come from the label claim instead of from cost. That is normal for this application — a carrageenan-free claim is usually bought, not saved into.
Stability is the part that actually decides it, and that is checkable before you run the trial: the formula pre-check flags the parameter combinations that usually break.
Three cases where you should not switch
The credibility of everything above depends on these.
- You need body and nothing else. A barista oat milk whose whole sensory proposition is a thick, fatty pour over coffee is a viscosity problem, not a suspension problem. Gellan will not fix it, and stacking gellan with enough body builder to match carrageenan can read as slimy. Keep the carrageenan, or move to a fibre-based system that was designed for body.
- Very high calcium fortification. Above roughly 0.3% added calcium, the calcium load that makes gellan set is also enough to make it set too early — in the line, before the filler. You can manage this with citrate, but the window gets narrow. Test the calcium level before you commit, and expect to give up some of the fortification.
- A protein level or base you have not run before. Pea and almond bases at low protein settle differently from oat, and the dose that suspends cocoa in oat may be too low for a mineral-fortified almond drink. Do not transfer a validated formula across bases without a settling test.
Migration in four steps
Each step has a pass criterion. Do not move to the next one until it is met — most failed projects skip step 2.
- Bench trial at the target dose. Hydrate gellan at 85–95 °C in the water phase, add calcium as CaCl₂ at 0.02–0.05% (or confirm the base already carries enough), add 0.03–0.06% sodium citrate to hold the set, then add the protein base. Pass: the mix is fluid above 60 °C and forms a weak gel at room temperature that breaks when stirred.
- Confirm hydration, visually. Dilute a sample 1:10 and look for undissolved particles under a light. Pass: no "fish eyes", no haze beyond the base's own. If it fails, the problem is temperature or dispersion, not dose — raise the hydration temperature and pre-blend the gum with a dry carrier.
- Ion and pH fine-tuning. Measure the set temperature with the sequestrant in place, then adjust citrate until the set happens after the filler. Pass: fill temperature at least 10 °C above the measured set point, with margin through the whole run.
- Stability. Hold samples at 4 °C, 20 °C and 30 °C for 4 weeks, and run one accelerated cycle. Check for sediment, a cream ring, and gel fragments. Pass: no sediment layer at the bottom, no gel lumps, transmittance within specification.
Failure modes
These are the combinations that most often produce a rejected batch on a first gellan trial in a plant milk.
| Symptom | Likely cause | Direction to fix |
|---|---|---|
| Product sets in the line or in the filler | Free calcium plus a slow pass through 30–50 °C with no sequestrant | Add 0.05–0.06% sodium citrate, or add the acid last if the system is acidified |
| Thickens slightly, never suspends, cocoa settles in a week | Dose below the network threshold, or soft water with no added calcium | Raise dose toward 0.03%; add 0.02–0.05% CaCl₂ |
| Suspends at day 1, sediment appears at week 6 | Hard water variability changing the network, or protein interaction at high protein | Test the water source across seasons; re-check dose against protein level |
| Reads thick and coating rather than refreshing | High acyl grade used at a suspension dose, or excess body builder | Move to low acyl and cut the builder |
| Haze or clarity loss against a clear-drink spec | Undissolved fraction, or dose at the top of the range | Confirm full hydration; test transmittance at the actual dose |
| Fouling on the heat exchanger | Protein and gum interacting at high protein plus high calcium | Reduce calcium before UHT, adjust homogenisation pressure, review hold time |
What to specify when you order
The grade matters more than the brand. Ask for low acyl gellan, gel strength in a defined range, and a COA that reports the parameters you actually control — not just a specification sheet. For a plant milk the relevant lines are gel strength, clarity (transmittance), loss on drying, and the microbiological limits for an ambient-filled product. If you are running a clean-label claim, confirm the carrier used in any pre-blended version, because a maltodextrin carrier will appear on the label.
The dose you validate on one lot has to hold on the next. Ask for lot-to-lot gel strength data before you scale, and design against the bottom of the range rather than the middle — a formula that only works on the strongest lot is a formula that will fail in production.
Tools
- Dosage calculator — dose, hydration temperature and ion requirement for 32 applications, including plant-based drinks at 0.015–0.040%.
- Formula pre-check — flags the hydration, pH and ion combinations that fail first.
- Troubleshooting tree — works back from the symptom you are seeing to the likely cause.
Frequently asked
Can I replace carrageenan with gellan at the same dose?
No. Gellan is effective at roughly half to two-thirds of the carrageenan dose for suspension, and it brings no body. Use 0.015–0.030% low acyl gellan against a carrageenan dose of 0.020–0.050%, and rebuild viscosity separately.
Does gellan change the taste?
At these doses it has no taste of its own. What changes is the mouthfeel: without a body builder the drink reads thinner. That is usually reported as a taste problem by a sensory panel even though it is a texture problem.
Will it survive UHT?
Low acyl gellan survives direct and indirect UHT. High acyl gellan is the grade that struggles through UHT and melts in retort, so do not substitute the grade when a supplier is out of stock.
Does the calcium in the base count?
Sometimes. Test the base water and the base itself before adding calcium — a mineral-fortified or hard-water base may already carry enough to set the network, and adding CaCl₂ on top will make it set too early.
Is gellan allowed in organic products?
Gellan gum is not on the EU organic inputs list for food additives in the same way as some other hydrocolloids, so an organic-certified plant milk needs its certifier's confirmation before you design around it. Check before the trial, not after.