Why RTD Tea Goes Cloudy: Diagnosing Tea Cream in Green Tea Beverages
From EGCG–caffeine complexation and calcium to the turbidity, pH and solubility numbers that belong on your purchase order.
Quick answer. Cloudiness in chilled green tea is tea cream: colloidal particles in which EGCG–caffeine complexes come out of solution as the infusion cools. Calcium and magnesium accelerate it, and a lower pH makes it worse. The diagnostic number is turbidity in NTU — but only when the measurement temperature and solids concentration are stated. Cold-water-soluble green tea extracts are typically specified below 50 NTU, against 100 NTU or more for hot-water-soluble grades.
1. Four symptoms, four different causes
"The beverage went cloudy" is not a diagnosis — it is a description of at least four distinct failures, each with a different fix. Before contacting a supplier, identify which one you have.
| Symptom | Likely mechanism | Check first | Lever |
|---|---|---|---|
| Haze at 4–7 °C that clears when rewarmed | Reversible tea cream: EGCG–caffeine and polysaccharide–protein colloids | Turbidity at 7 °C, then again at 20 °C | Cold-water-soluble grade; turbidity ≤50 NTU |
| Sediment after 2–4 weeks that will not redissolve | Cream particles matured past redissolution, or calcium salts of organic acids | Calcium and magnesium in process water | Demineralised water; calcium removal |
| Progressive darkening after UHT or retort | Heat-driven catechin oxidation, metal-ion catalysed | Iron and copper in water; dissolved oxygen | Low-metal water; chelation or ascorbic acid |
| Cloudy immediately on make-up, at any temperature | The powder itself is not cold-water soluble | Turbidity of a 1–2% solution made in cold water | Specify CWS, not HWS |
Typical commercial CWS green tea extract specifications: turbidity <50 NTU, pH 5.2–5.8, bulk density 0.50–0.60 g/cc, moisture ≤5%. Hot-water-soluble grades of the same material are commonly specified at <100 NTU.
2. What is actually in the precipitate
A knock-out/knock-in study of green tea cream (published in Foods, 2023) identified polysaccharides, proteins, EGCG and caffeine as the components that participate, in the form of polysaccharide–protein and EGCG–caffeine colloidal particles. Critically, the interaction of EGCG or caffeine with the macromolecules alone was not the driver — the EGCG–caffeine complex itself was. That is why reducing either EGCG or caffeine concentration is a legitimate control lever, and why a high-EGCG, high-caffeine specification is the hardest combination to keep clear.
Analysed cream from green tea infusions runs roughly 30–79% polyphenols, 15–28% total sugars and 2–10% caffeine, with catechins making up 13–43% of the polyphenolic fraction. Green tea hazes rather than creams to the same degree as black tea, because it lacks the theaflavins that initiate creaming in black tea infusions.
Two formulation variables dominate. Water hardness: calcium, magnesium, iron and zinc all enhance cream formation in green tea infusions, with calcium combining with organic acids — oxalic, quinic and tartaric — to drive sediment. pH: lowering pH decreases polyphenol solubility, shifts the phase boundary and produces larger particles, so an aggressively acidic formulation will cloud more readily, not less.
Concentration matters as much as temperature. Below roughly 2% w/w tea solids, cream forms mainly by the growth of discrete complexes; above that, demixing by spinodal decomposition becomes the dominant mechanism, driven simply by the limited solubility of the polyphenol fraction. This is why a trial that looks clean at serving strength can fail in the concentrate, and why turbidity should be evaluated at both make-up and concentrate strength before a formulation is frozen.
3. How suppliers make an extract cold-water soluble
Three routes are in commercial use, and they leave different fingerprints on your COA. Tannase treatment cleaves the gallate ester bond on catechins, releasing free gallic acid, which competes with the larger polyphenols for caffeine and yields smaller, more water-soluble complexes. Alkali solubilisation with sodium or potassium hydroxide breaks the hydrogen bonds between polyphenols and caffeine. Physical decreaming — chilling the extract and removing the precipitate by centrifugation or filtration — removes cream-forming material rather than chemically modifying it.
The practical consequence for a buyer is that two extracts can carry an identical EGCG assay and behave completely differently in a chilled bottle. Tannase treatment changes the catechin profile, decreaming removes a fraction of it, and alkali treatment leaves a sodium or potassium load. Ask which route was used, and assay the finished lot rather than the input material. A fourth option — adding a solubiliser such as a sugar or polyol — lowers measured turbidity without removing the cream-forming material at all, which is worth knowing before you accept a clear sample as proof of a clear product.
4. What to write on the purchase order
Write turbidity as a complete method, not a bare number. State the NTU limit, the solids concentration, the temperature, and the instrument method — the nephelometric method referenced in the relevant patent literature measures at refrigeration temperature, about 7 °C, because that is where the failure appears. A 50 NTU result measured at 20 °C is not comparable to one measured at 7 °C, and this is the most common source of dispute between a beverage formulator and an ingredient supplier.
As a reference point, purified green tea extracts assessed as clear in the patent literature come in at roughly 7–32 NTU, while failed comparatives sit at 178–314 NTU. Pair the turbidity limit with pH (5.2–5.8 for typical CWS green tea), bulk density, moisture, mineral content and a defined cold-water solubility test — "soluble" without a temperature, concentration and time is unenforceable. Finally, set numeric limits on calcium and magnesium in the extract, since those ions will work against you regardless of how good the extract is.
Add two shelf-life conditions to the same clause. A freeze–thaw cycle is the standard accelerated stress for a chilled distribution chain: a concentrate that survives one cycle at 7 °C may throw sediment on the third, and a supplier specification that only certifies the powder at make-up tells you nothing about that behaviour. And because the failure is time-dependent, require a stated turbidity limit at a defined endpoint — for example, still under 50 NTU after 14 days at 4 °C — rather than a single measurement on the day of production.
