How dessert icing crystallizes in the kitchen and the lab

Despite icing’s popularity in desserts, scientists haven’t conducted many studies on it and how it crystallizes. Many papers on the topic focus on industrial settings and the intentional seeding of a solution with sugar crystals.

But bakers can make icing without seeding it at all: After boiling a deceptively simple mixture of water and sucrose, they quickly cool the resulting syrup, and once the temperature has dropped enough, they vigorously stir the mixture. A nucleation process then rapidly forms many tiny crystals, giving the icing—called poured fondant—a smooth, creamy texture.

To learn more about the formation and physical behavior of fondant, Hannah Hartge and Thomas Vilgis , of the Max Planck Institute for Polymer Research in Mainz, Germany, and Eckhard Flöter, from the Technical University of Berlin, made the food in the lab. The team found that the mixture undergoes a multistep process that first produces a broad size distribution of crystals before they’re smashed into a narrow range of smaller-sized ones.

More sucrose molecules can dissolve into a volume of water at temperatures around water’s boiling point than at a lower temperature. The supersaturated, metastable system can be modeled with classical nucleation theory to describe its crystallization. But the high mechanical energy in the system makes it difficult to model and limits the inferences that scientists can draw from simple phase diagrams of sugar.

Hartge and her colleagues used light microscopy to analyze the solids formed at various times. To induce crystallization, they used a mechanical kneading machine to apply high shearing forces to the mixture. In the supersaturated solution, they found that the torque over time experienced a decrease immediately followed by a sharp increase before leveling off again, seen in the figure below.

The torque initially agitates the solution and promotes fast crystallization. Later, the growing crystals hit each other because of shearing forces. The minimum in torque is caused by a flurry of crystallization as the sucrose concentration drops in the remaining liquid. The crystal sizes are largest (tens of micrometers) at the peak. At that point, the crystals become so big (marked by black arrows) that the largest conglomerates collide with each other and break apart. But smaller fragments of 1–5 µm in size continue to grow and develop to produce the uniform icing.

Previous efforts to model the classical nucleation of agitated crystals have approximated the shear rate as constant. That assumption, however, fails to capture the complex behavior of a kneading machine. The new result by Hartge and colleagues shows that at least for supersaturated sugar solutions, crystal nucleation can be modeled classically even though the shear rate induced by the kneader is strongly time dependent. The findings, therefore, may help improve the understanding of the physics of sucrose crystallization under such highly nonequilibrium conditions. (H. M. Hartge, E. Flöter, T. A. Vilgis, Phys. Fluids 35, 064120, 2023 .)