Why Acidic Fruit Curdles Dairy Gelato, and How to Stop It

Milk sits at pH 6.7. Passion fruit sits near 2.9. Pour one into the other and you have not made a fruit gelato base, you have started making fresh cheese. The graininess in the tub is casein doing exactly what casein is supposed to do.

Smooth is the exception here, not the default. A fruit dairy base only stays smooth if you keep it away from one specific pH.

The number that explains the whole defect

Casein is not dissolved in milk. It is suspended as micelles, colloidal particles held apart by a brush of κ-casein on their surface and by a shared negative charge, with colloidal calcium phosphate holding the interior together.

Drop the pH and both of those supports fail at once. The negative charge falls away as acid neutralises it, and the calcium phosphate dissolves out of the micelle interior. At pH 4.6, the isoelectric point of casein, the net charge reaches zero, nothing keeps the particles apart, and they aggregate into a curd. This is not a defect of your recipe. It is the reaction that yoghurt, ricotta and every fresh acid-set cheese is built on, and casein does it whether or not you wanted it to.

So the working rule is simple and numeric. The dairy phase has to stay meaningfully above 4.6, and in practice you want a floor around 5.2, because the mix is not uniformly mixed at the moment the purée hits it. Local pockets reach the isoelectric point long before the bulk reading does.

Quick reference. Casein aggregates at pH 4.6. Keep a dairy fruit mix above about 5.2, add fruit cold, and never reheat an acidified mix. Heat plus acid together is far more destructive than either alone.

Figure 1 — Most of the fruit worth putting in a gelato sits well below the pH at which casein gives up.

It is worth being precise about why heat makes acid so much worse, because it is not simply that hot things react faster. Heating milk past roughly 70 degrees denatures beta-lactoglobulin, the main whey protein, and the denatured protein then binds to the kappa-casein layer on the micelle surface through disulfide bonds. That modified surface aggregates at a higher pH than a native one does, which is precisely why yoghurt makers heat their milk before culturing it. In a gelato kitchen the same chemistry works against you: a mix that would have survived at pH 5.0 cold can fail at 5.3 once it has been through the pasteuriser with the fruit already in it.

Which fruits are actually dangerous

Approximate pH values for common fruits are published in the FDA's food pH tables, used for acidified-foods classification. The ranges vary with cultivar, ripeness and season, so treat them as a triage tool rather than a specification, and confirm with a pH meter on your own lot.

Two things get missed reading a table like this. The first is that a titratable acidity reading tells you how much buffering capacity the fruit brings, which predicts how far it will actually pull your mix, while pH alone only tells you where the fruit starts. The second is that any added acid in the recipe stacks on top of the fruit's own.

The protease trap, which is a different failure

Fresh pineapple, papaya, kiwi and fig carry active proteases: bromelain, papain, actinidin and ficin respectively. These do not curdle the mix. They hydrolyse the casein directly, and the symptom is different and often misread: a base that goes thin rather than grainy, sets poorly, and can develop a bitter edge from the peptides released.

The fix is the opposite of the acid advice. Proteases are denatured by heat, so a brief cook of the fruit purée solves it completely, which is why canned pineapple behaves and fresh pineapple does not. If you have a fruit that is both acidic and proteolytic, cook the purée to kill the enzyme, cool it fully, and only then combine it with the dairy.

Four levers that actually work

Add cold, add last. Pasteurise the white base alone, chill it, and fold the purée in at maturation. This costs nothing and removes the heat-plus-acid combination that does most of the damage. The same discipline that protects pigment protects protein.

Buffer with citrate. Sodium citrate is the workhorse here. It buffers the mix against the pH drop and sequesters calcium, which loosens the ionic bridging that drives aggregation. Small additions, in the region of 0.1 to 0.3 per cent, move the failure threshold measurably without a flavour penalty.

Use a stabiliser designed for acid milk. High-methoxyl pectin is the standard protective colloid for acidified dairy: it adsorbs onto the casein surface at low pH and provides steric stabilisation, which is exactly how commercial drinking yoghurts survive at pH 4. Propylene glycol alginate does similar work. Note that both need proper hydration before the acid arrives, so follow the hydration temperature guide and get them into the base during the hot stage.

Change the architecture instead of fighting it. For lemon, lime and passion fruit, the honest answer is usually that they do not belong in a milk base at all. Make them as a sorbetto, or keep the acid out of the base entirely and deliver it as a variegate ribbon folded in at the end, where it never has time or contact to destabilise the surrounding dairy.

Reading the damage you already have

Not every acid problem announces itself as visible curd. Marginal destabilisation is more common and harder to diagnose.

The tell is a slightly sandy or chalky mouthfeel with no obvious ice, combined with a faster meltdown and a wetter puddle than the recipe should give. That is aggregated protein that never grew large enough to see, sitting in a mix whose emulsion no longer holds water the way it did. It reads as a texture fault, so people usually chase it as an icy-texture problem or as graininess and adjust solids, which changes nothing because the cause is upstream.

Diagnose it by measuring the pH of the finished mix, not the fruit. If the mix reads below about 5.0, you have your answer, and the fix is buffering or a change in process order rather than another point of total solids. A protein-concentrate addition, often reached for when a base feels weak, makes an acid problem worse rather than better, because it adds more of the thing that is aggregating.

One last habit worth building: take the reading after maturation, not only at the moment of mixing. Fruit purées carry their own microflora and their own slow chemistry, and a base that measured 5.3 when it went into the fridge can read below 5.0 the next morning. If a recipe sits close to the floor, either shorten the maturation, buffer it properly, or accept that this particular fruit belongs somewhere other than a milk base.

Related Concepts

• Sorbet separation is the dairy-free cousin of this defect, driven by phase behaviour rather than protein.
• Passion fruit sorbetto shows the usual destination for the fruits at the sharp end of the table.