A bakery fruit filling is a harder problem than it looks: it has to be pumpable when cold-applied, hold shape immediately after deposition, survive oven temperatures without boiling out, withstand freeze-thaw if the product is frozen, and not weep water into the pastry beneath it. Agar does several of those reasonably and one of them badly. Low acyl gellan at 0.05–0.15% covers a wider part of that list, and it does so at roughly a tenth of the agar weight, which changes another property people notice — the flavour comes through instead of being muted by the gum.
What agar does, and where it stops
| Requirement | Agar at 0.5–1.0% | Low acyl gellan at 0.05–0.15% |
|---|---|---|
| Sets firm enough to pipe or deposit | Yes | Yes |
| Holds shape through bake-off | Softens and can run | Holds |
| Freeze–thaw stability | Poor — syneresis after one cycle | Good |
| Water binding over shelf life | Moderate weeping | Low weeping |
| Flavour release | Gum can mute fruit notes | Cleaner at this dose |
| Clarity | Translucent | Clearer, matters for glazes |
| Dose to reach set | Higher | roughly a tenth |
The freeze–thaw row is the one that generates complaints. Agar gels are prone to syneresis — the network contracts and expels water — and a frozen filled product put through one thaw cycle often arrives with water sitting on top of the filling. That is a structural property of the agar network rather than a formulation mistake.
The substitution table
| Line | Agar filling | Gellan filling | Difference |
|---|---|---|---|
| Dose | 0.5–1.0% | 0.05–0.15% | roughly one tenth |
| Powder per 1,000 kg | 5–10 kg | 0.5–1.5 kg | −4 to −8.5 kg |
| Setting requirement | Cool below ~40 °C | Ions present, usually from the fruit | Fruit calcium often sufficient |
| Bake-off behaviour | Can soften and run | Holds shape | Relevant for bake-stable lines |
| Freeze–thaw | One to two cycles before weeping | Several cycles typically | Test your own cycle count |
| Flavour impact | Noticeable above mid range | Minimal at this dose | Sensory panel, not instrument |
Read the setting row carefully. Low acyl gellan needs divalent cations to crosslink, and fruit preparations usually contain enough calcium on their own. That is convenient and also a trap: change fruit supplier, change fruit solids, or move to a different water source, and the set moves with it. Treat the cation content like a raw material specification rather than a background condition.
The dose arithmetic
| Line | Agar filling | Gellan filling |
|---|---|---|
| Powder weight per tonne | 5–10 kg | 0.5–1.5 kg |
| Weight ratio | 1 | roughly 1/10 |
| Additional salts sometimes needed | No | Small, if the fruit is low in calcium |
| Break-even consideration | — | Below roughly ten times the agar price per kilogram, less material goes into the filling |
Both columns above are weights, which is the part that is a property of the formulation. The price multiplier at which this stops being a reduction is about ten — above ten times the agar kilo price you are paying for performance rather than saving material, and for bake-stable or freeze-thaw products that is still usually the right trade.
Three cases where you should not switch
- Low-solids systems with nothing to carry. At low solids the water is not bound by anything else, and neither gum will save it. Fix the solids before blaming the gum.
- Products sold on agar's specific firm snap. Some filled products rely on a brittleness that agar gives and gellan does not. If that texture is the product, keep it.
- Very high calcium fillings you are not going to control. With abundant calcium the gellan window narrows and overdosing gives a rubbery filling. Either control the cations or stay with agar.
- Immediate cost-driven reformulation with no performance problem. If the current agar filling is not failing, this change buys you very little.
Migration in four steps
- Write down which requirement is failing. "Weeps after freeze-thaw" and "runs during bake-off" lead to different solutions. If nothing is failing, stop.
- Check the calcium content of the fruit before setting a dose. This single measurement determines most of the outcome, and it varies by supplier and season.
- Dose at the bottom, 0.05%, and work up. The window is narrow, and over-dosing gives a rubbery rather than firm result.
- Run the real cycle. Two freeze-thaw cycles minimum, plus bake-off at the actual line temperature. Bench set is not the product.
Failure modes
| Symptom | Likely cause | First check |
|---|---|---|
| Water on the surface after freezing | Insufficient network strength | Increase dose slightly, then check solids |
| Filling runs during bake | No thermostable component | Confirm it is low acyl gellan and not high acyl |
| Rubbery rather than firm | Dose above the useful range | Drop to the bottom of the range |
| Grainy set | Incomplete hydration | Dispersion and hydration temperature |
| Set varies between fruit lots | Calcium variation between suppliers | Analyse incoming fruit solids and calcium |
| Flavour muted | Dose too high for the fruit | Lower dose; this is usually fixable |
Tools
- Dosage calculator — includes bakery fruit fillings at 0.05–0.15% and fruit preparations at 0.03–0.10%.
- Troubleshooting tree — start from the defect, work back to the parameter.
- Formula pre-check — flags ion and pH combinations that fail first.
Frequently asked
How much gellan replaces agar in a bakery filling?
Roughly a tenth of the weight — 0.05–0.15% against 0.5–1.0% of agar. Because gellan needs calcium to set and fruit supplies much of it, confirm the fruit's calcium content before fixing the dose.
Will it survive freeze-thaw?
Typically several cycles where agar manages one to two before weeping. Test with your own cycle: number of cycles, temperature and duration all matter more than the nominal figure.
Does the fruit supply enough calcium?
Often yes, which is why the dose is low. It is also the main source of variation — fruit suppliers and seasons differ, so analyse rather than assume.
Does the filling taste different?
Usually cleaner. At roughly a tenth of the gum weight there is less material muting the fruit, which most panels register as a stronger fruit note rather than as a texture change. Check it with people rather than instruments.