A pinch of flour prevents the formation of protein aggregates, which nonetheless can be broken up in a mixer.
HOW CAN ONE SAVE A CRÈME ANGLAISE that has curdled? The question is of culinary importance because crème anglaise figures in one form or another in many desserts. One of the differences between crème anglaise and crème patissière, for example, is that although both are composed of milk, egg yolks, sugar, and an aromatic ingredient such as vanilla, crème patissière additionally contains a certain amount of flour, which protects against curdling. Crème anglaise, lacking this protective agent, is more liable to turn.
The Second International Workshop of Molecular Gastronomy, held in April 1995 at the Ettore Majorana Center in Sicily, brought chefs and physical chemists together to examine such questions. Two renowned chefs, Christian Conticini and Raymond Blanc, raised a series of issues relating to sauces and dishes derived from them: crème anglaise, crème patissière, mayonnaise, stiffened egg whites, soufflés, chocolate cream fillings, jellies, preserves, and so on. The physicists and chemists, led by Nicholas Kurti, Pierre-Gilles de Gennes of the École Supérieure de Physique et de Chimie Industrielles in Paris, and me, sought to identify the mechanisms for culinary effects that were plain to see but not yet understood scientifically, regarding sauces as solutions, emulsions, foams, gels, suspensions, and so on.
At the conference the problem of crème anglaise was examined experimentally. Like many other delicate dishes, crème anglaise has inspired many dictums and ad hoc remedies that are worth scrutinizing. It has long been said that a pinch of flour added to crème anglaise will keep it from curdling. Why should this be so? Is it true that a crème anglaise that has turned can be saved by vigorously shaking it in a blender?
The Pinch of Flour and the Blender
Crème anglaise was examined at different stages of preparation. First, the custard was gently heated (at a lower temperature of 65°C [149°F]); gradually, as in the case of every successful crème anglaise, the mixture of milk, sugar, and egg yolks thickens. Under the microscope, small-scale structures (a few micrometers long) can be seen.
Then this same custard was put in a microwave oven for a few seconds so that a curd—a sign of overcooking—would appear. The observed microscopic structures were about twice as large and dense as those that had been observed in the successful custard, but their general aspect was not substantially different. When the custard was overheated, its appearance under the microscope changed completely: Clear liquid areas separated very dense areas composed of structures similar to those that had been observed at the onset of curdling.
Finally, this botched crème anglaise was mixed for a few dozen seconds. To the naked eye it had become frothy. The curdles had disappeared, and the texture of a perfect custard seemed to have been restored; under the microscope, however, a state of aggregation intermediate between that of a perfect custard and that of the initial stage of curdling could be discerned.
Inevitable Coagulation
It seems clear that the setting of a crème anglaise depends on the coagulation of the egg yolk, which occurs whether the result is successful or not. The structures observed under the microscope probably are aggregates of proteins that have been partially broken down by the heat, then reassembled by means of weak chemical bonds.
When the crème anglaise has been excessively heated, coagulation is rapid and produces macroscopic aggregates, evidence of curdling, which can be eliminated by mixing. Has the possibility of curdling thereby been completely removed? Can a perfect crème anglaise be reconstituted by putting a botched one in a blender?
Microscopic analysis shows that although a blender does a good job of dissociating the macroscopic aggregates, more agitation is needed if the result is to contain only microscopic protein aggregates, similar to the ones found in a successful sauce. Those whose palate is sufficiently refined to detect the omelet taste characteristic of a botched crème anglaise can prevent the sauce from curdling in the first place by adding a pinch of flour before cooking. Curdling will not take place, even if the crème anglaise is boiled.
The reasons for this protection are still a subject of debate, but it is known that placing starch granules in a hot liquid triggers the release of amylose molecules and that the water penetrates the granules and causes them to swell. These swollen granules, together with the long, dissolved amylose molecules, limit the movement of proteins, blocking the formation of macroscopic protein aggregates.