Why Milk-Chocolate Crémeux Becomes Grainy, Greasy or Too Soft

This article diagnoses the published Callebaut milk-chocolate crémeux used as the reference for the Chocolate Tonka & Passion Fruit Cake. It is not a claim that this is Tanya Novak's proprietary formula.

When a crémeux goes wrong, I do not begin by adding more chocolate or gelatin. I ask one question first: when did the problem appear? A grainy custard in the saucepan, an oily emulsion in the jug and a soft insert after thawing may all be described as a bad crémeux, but they have different causes and different chances of rescue.

I keep a small sample at each stage when I am testing a new batch. That turns a vague disappointment after thawing into a much shorter investigation.

I first look at the custard before it touches the chocolate. If I can already see tiny cooked particles, the yolks have coagulated too far—usually because the heat was too high, a corner of the pan was not scraped, or residual heat continued cooking the custard. For this formula, Callebaut specifies 82°C / 180°F. I stop there and strain immediately.

A fine sieve can remove a few particles caught early. It cannot make severely scrambled custard new again. Blending may make the particles less visible, but the protein structure has still been damaged and the final texture may remain sandy.

If the custard was smooth, I inspect the grains. Chocolate fragments soften with gentle warmth; cooked egg does not melt. Gelatin strings feel elastic. This simple observation prevents me from treating three different materials as one fault.

A greasy film, clear beads of fat or a dull broken surface usually means the aqueous custard and the chocolate fat have not formed a fine emulsion. This often happens when the chocolate is too cold, the custard is added without establishing an elastic centre, or the mixture cools below its useful working range before the emulsion is complete.

L’École Valrhona advises keeping the developing chocolate crémeux above approximately 35°C / 95°F so most cocoa butter remains melted. I add the custard progressively, stir from the centre until it becomes glossy and elastic, then finish with the immersion blender fully submerged.

If the ingredients are still safe and the split has just happened, gentle warming and controlled submerged blending may bring the emulsion back. I do not use strong direct heat, and I do not pour in extra cream as a reflex. Both actions can replace one problem with a new formula.

Warm crémeux is supposed to be fluid. I judge the set only after the specified cooling or freezing cycle and at the temperature at which the cake will be served. A centre tasted half-frozen can seem firm; the same centre may lose its shape after a complete refrigerator thaw.

For this formula, the first check is notation. Thirty-two grams of prepared gelatin mass is not 32 g dry gelatin, and it is not automatically interchangeable with 32 g of hydrated sheets. I record the dry gelatin, hydration ratio and Bloom strength so the same words describe the same material in every batch.

Next I confirm the couverture. Milk chocolates with the same cocoa percentage can contain different amounts of cocoa butter, non-fat cocoa solids, milk solids and sugar. A gram-for-gram product change can therefore alter both warm fluidity and cold structure.

I also check whether the anglaise actually reached 82°C / 180°F, whether every ingredient was weighed correctly, whether the gelatin dissolved evenly, and whether the insert received enough uninterrupted time to set or freeze completely. I change only one variable in the next test.

Adding gelatin can make a soft insert hold, but it may also create a rubbery layer that no longer melts with the mousse. Adding chocolate can increase structure while also changing sweetness, cocoa intensity and the balance with passion fruit. Neither proves what failed.

My smallest useful test is a half batch divided between two identical shallow containers. I keep one as the control and change one variable in the other—for example the verified gelatin mass or the original specified couverture. Both samples go through the same chilling, freezing and thawing schedule.

In the Chocolate Tonka & Passion Fruit Cake, the milk-chocolate crémeux must stay smooth and precise beside a sharper mango–passion fruit gel and a lighter dark-chocolate tonka mousse. A greasy centre makes the whole cake feel heavy; a weak centre blurs the cut; a rubbery correction interrupts the melt.

This problem page therefore connects the finished cake to the reference crémeux formula, crème anglaise, chocolate emulsification, gelatin mass, milk-couverture selection and freeze–thaw stability. Each connection answers a different diagnostic question instead of repeating the same recipe in several articles.

Grainy usually sends me back to custard heat or undissolved particles. Greasy sends me to the emulsion and its temperature. Too soft sends me to gelatin notation, chocolate identity, accurate weights and the full setting and thawing cycle.

The time at which the defect first appears is often more useful than the final appearance. Find that moment before changing the formula.