Croissants are literally the stuff of legend. One often-told story is that the pastry was created by Viennese bakers after the city defeated the Turkish army at the Battle of Vienna in 1683. The crescent was a prominent part of the flag of the Ottoman Empire, so by enjoying a croissant you could symbolically bite the vanquished enemy. Other stories trace the croissant back to the Battle of Tours in 732, when Frankish and Burgundian forces defeated an army of the Umayyad Caliphate.
You don’t need much to bake croissants, just some dough, chiefly made from wheat flour and water, and some fat, typically butter. In large-scale manufacturing, sheets of the dough and fat about 1 cm thick are layered on top of each other or the fat is extruded between dough sheets. Repeated rolling and folding, a process called lamination, creates alternating layers of dough and fat each about 100–200 µm thick. The mutually alternating layers, typically 18–32 in a croissant, give the bread its billowy appearance and flaky texture: During baking, the vaporization of water in the dough inflates the layers, which expand like balloons.
Not too hard, not too soft
The rheological properties of the fat—that is, the way it deforms and flows—have a significant bearing on the quality of the croissant. A fat has to have just the right degree of hardness for lamination to succeed. A fat that is too hard can break during lamination and can also rupture the dough. A fat that is too soft will absorb into the dough. So the wrong fat can translate into dense, crumbly, soulless croissants and unhappy customers. Luckily, bakers and scientists, long concerned with textural and mouthfeel food attributes (see the Quick Study by Erich Windhab, Physics Today, June 2006, page 82), have recognized that not every fat is appropriate for every task. Thus they have contributed, from empirical experience, to the development of fats for specialized functions.
An example of a type of fat appropriate for a specific purpose is the so-called roll-in or puff-pastry fat used in croissants. The best croissants are made with butter; indeed, consumers are desirous enough of butter that some regions of Europe are currently experiencing a butter shortage. Sometimes, though, constraints of cost or processing temperature dictate that other roll-in fats be used. In our lab work, for example, we studied croissants made with shortening. The different roll-in fats have somewhat different properties. For instance, to achieve good lamination with butter, the temperature needs to be held in the relatively narrow range of 15–20 °C, as bakers well know. Shortening will laminate well over a wider temperature range.
The rheological response of roll-in fats arises from an underlying colloidal network of fat crystals. As discussed below, the network, sculpted by the fat’s formulation and crystallization, has structure on multiple scales. To ensure good lamination characteristics, the crystallization must be carefully carried out. The roll-in fats are cooled and mechanically worked so as to produce many small crystals and to impart a degree of microscopic ordering to the crystal network.
The professional baker evaluates a roll-in fat by pressing and manipulating it between the fingers and relying on expert feel. For the scientist, a more appropriate approach might be to subject fats to a well-defined deformation—for example, a periodic oscillation—and to express their behaviors in a stress (internal force per unit area) versus strain (relative extension) plot, such as panel a of the figure.
The technique we used to obtain the plot is called large amplitude oscillatory shear, or LAOS, rheology. The approach is sufficiently new that it is itself a topic of research, but in any case, it can be used to probe the structure and function of various materials, including polymers, colloids, and foods. LAOS decomposes a material’s oscillatory response into elastic and viscous contributions.
As panel a shows, for very small strain, the stress of both roll-in and all-purpose fat responds linearly. The fats act like viscoelastic solids and deform reversibly; in fact, that solid-like response is observed across the frequency spectrum. The linear regime is very narrow and terminates at a critical strain on the order of 0.01%, a value typical of materials acting via short-range or van der Waals interactions. Strains beyond the critical value are sufficient to disrupt the network and initiate irreversible yielding behavior. During the lamination process, a roll-in fat must withstand pressures that are easily sufficient to drive the material to the nonlinear regime.
Our LAOS experiments reveal that the two types of fat gave comparable maximum stress, or yield stress, in the range of 4000–5000 Pa. But their yielding behaviors are quite different: Roll-in fat displays a broad stress plateau when deformed, whereas the all-purpose fat shows an abrupt stress drop indicative of internal breakage; thus an all-purpose fat cannot laminate well.
The rheology of fat arises from the hierarchical self-organization of its structure, which combines physical length scales ranging from a few angstroms to several tens of micrometers. At the molecular level, fats are made of triglyceride (TG) molecules in which a trio of fatty acids emanate from a glycerol backbone. As much as we love croissants, we would be remiss if we did not note that the roll-in fats so essential for their texture may have a high content of trans-fatty acid, whose excessive consumption is detrimental to cardiovascular health. Indeed, one of the motivations for our research is to devise alternatives to roll-in fats that are healthier but yield equally delicious and flaky croissants.
X-ray scattering experiments distinguish three structural levels in roll-in fats, as reflected in the distinct power-law regimes evident in a plot of the ultra-small-angle scattering intensity versus scattering angle. The first, smallest-scale level encompasses nanosized platelets, TG crystals (see panel b in the figure). Those platelets are smaller in roll-in fats than in all-purpose fats and have smoother boundaries. The second level, which is absent in all-purpose fats, comprises crystal aggregates, cylindrical in form and with modest height-to-diameter ratios, that spontaneously stick together to form clusters with submicron and micron sizes. It is those clusters that make up the third structural level, a microscopic crystal network in which liquid oil is embedded.
The complementary technique of electron microscopy reveals that in roll-in fats, the crystal clusters are organized into an ordered, layered network, as shown in panel c of the figure. The regular fats are, by comparison, relatively disordered.
The different length scales in a fat’s structural hierarchy have important implications in energy dissipation, as observed in their rheology. In biological and natural materials, the incorporation of multiple length scales can confer toughening and defect tolerance. The reason is that the scale hierarchy, in conjunction with a suitable microscopic mechanism, enables the local dissipation of high stresses that would otherwise induce material failure. Indeed, roll-in fats, with their tripartite structure, dissipate about 10 times more stress energy than do all-purpose fats. Stress dissipation in fats and the response of their microstructure under nonlinear deformations are topics that still warrant investigation. One intriguing issue is the possibility of an additional feature of roll-in fats: a sliding of the crystalline layers that may also affect energy dissipation and croissant flakiness.
Braulio Macias Rodriquez is a postdoctoral fellow and Alejandro Marangoni (www.crcfoodandhealth.com) is a professor and Canada Research Chair in the department of food science at the University of Guelph in Ontario, Canada.