The enemies of successful sauces form because water diffuses slowly in starch paste and gelatin.
LUMPS—THE SHAME AND DESPAIR of the cook. Cookbooks suggest various ways to avoid them in making béchamel and other sauces that are thickened with flour. Some authorities say to make a roux first, cooking butter and flour and then adding milk (cold according to some, hot according to others); others insist on the opposite method, namely pouring the roux into the milk, which, again depending on the author, should be either cold or hot.
Which method is better? If we are willing to waste a bit of flour, butter, and milk in order to test the four possibilities we will find that the formation of lumps is at bottom a question of speed: When the roux is added gradually to the milk or the milk to the roux, lumps do not appear; however, they do form when the two ingredients are mixed together at once, especially when the roux is poured into boiling milk.
This experiment gives us a method, but it does not explain the phenomenon. Let’s study the matter further by simplifying it. The butter serves chiefly to cook the flour, eliminating its bland taste by creating odorant and taste molecules (the browning is associated with the chemical reactions that form these molecules), but it is not the cause of the lumpiness, which also results from combining flour and water. Why? Flour is composed principally of starch granules, themselves made up of amylose and amylopectin. These two types of molecule are both polymers, which is to say chains of glucose molecules (linear in the case of amylose, ramified in the case of amylopectin). Amylopectin is insoluble in water, but amylose is soluble in hot water. When the starch granules are put into hot water, they lose their amylose molecules, and the water fills the space between the amylopectin molecules, causing the granules to swell up and form a gel known as starch paste.
This description gives us a handle on the problem of lumps, for it suggests that flour deposited in hot water is quickly enveloped by a gelatinized layer that limits the diffusion of water toward the dry central core of the lump. But it is inadequate as an explanation, however, because it fails to explain why, if the water diffuses into the periphery of the lumps, it does not ultimately reach the center.
One-Dimensional Lumps
As we have seen, everything is a question of speed. To measure the rate of gelatinization, let’s simplify the matter further by making a one-dimensional lump whose center we can observe. Put some flour in a test tube and then pour water over it (if the water is colored one can follow its diffusion into the flour by observing the movement of a distinct boundary). At room temperature the water infiltrates the upper layer of the flour fairly rapidly at first but then penetrates further very slowly, by less than a millimeter an hour. Heating the water causes the granules to swell, so contact with this gelatinized region limits the advance of the colored boundary. In this case the rate of diffusion rises to several millimeters per hour.
It follows, then, that the center of the lumps remains dry because of the slowness of the water’s diffusion through the gelatinized periphery. When the water has diffused and swollen the starch granules, binding them together, they form a layer that retards further diffusion toward the center to such a degree that it remains dry (a phenomenon characterized more precisely by specifying the many molecular interactions that take place between the starch granules and the water). In other words, placing a quantity of flour measuring a centimeter in diameter in hot water causes it to become moistened to a depth of one or two millimeters from the periphery, with the center remaining dry longer than the time needed for most culinary purposes.
How can we get rid of these lumps? No fundamental physics is needed to solve the problem. It suffices to break up the lumps—with a whisk, for example—into particles smaller than the thickness of the starchy layer.
Soaking Gelatin
Does this theory of the formation of flour lumps apply to other types of lumps? Cooks know that leaves of gelatin must soak in cold water before being used in a hot liquid. Failure to observe this rule creates strings that are as difficult to eliminate as lumps of flour in hot water. Are these strings likewise composed of a dry center and a sheath in which the water and gelatin molecules are mixed? To find out, first measure the water’s rate of diffusion in cold gelatin by placing a small amount of coffee grounds on its surface. A hemispherical colored zone extends outward from this point at a rate of only about a centimeter a day.
Next, let’s repeat the experiment we conducted with the flour, only this time replacing the flour with gelatin. One then observes that the boundary of colored water sinks into the gelatin only very slowly. But a new phenomenon now appears: Under the layer of gelatinous solution, the gelatin that is untouched by the water has melted like butter.
Similarly, in a hot broth, a sheet of gelatin that has not been soaked beforehand has trouble dissolving and conserves a solid central part that melts from the heat. Its molecules stick to one another, forming the dreaded strings. In a sheet that has been soaked long enough to allow the water to gradually penetrate into the center, on the other hand, heating causes the gelatin to dissolve without strings.