When triglycerides are cooled from the melt to a temperature below their melting point, they undergo a liquid-solid transformation to form primary crystals with characteristic polymorphism and polytypism. These crystals associate, or grow into each other, to form increasingly larger aggregates, which further interact, resulting in a continuous three-dimensional network. The macroscopic properties of the resulting fat crystal network are affected by all these levels of structure, however, most directly by the level of structure closest to the macroscopic world. Most of the research in the past has been directed towards establishing a link between triglyceride structure and phase behavior, crystal habit, in particular crystal polymorphism, and the macroscopic properties of fats. In order to truly understand, and eventually predict, the macroscopic properties of fat-containing products, it is necessary to characterize and define the different levels of structure present in the material and their respective relationship to a macroscopic property. A macroscopic property is not always simply and directly related to molecular structure.

Work in our laboratory has shown that the rheological properties of fats are strongly dependent on the microstructure of their fat crystal networks using fractal scaling relationships and developed a novel particle counting method for the determination of the fractal dimension from polarized light micrographs (Narine and Marangoni, 1999. Physical Review E 59:1908; Narine and Marangoni,1999. Physical Review E 60:6991). We have also built a mechanical and structural model which relates the Young's modulus () of such soft material to particle properties, amount of solids, and the spatial distribution of the solid mass (Marangoni, 2000. Physical Review B 62, In Press;), namely:
           A             _1/3-D
e_s = ------  F
         pad²_o
where A is Hamacker's constant, a is the diameter of the primary crystal aggregates (~ 1-10µm), d_o is the distance between flocs of primary crystal aggregates,F is the volume fraction of solids, and D is the mass fractal dimension of the network. Using our model, it is possible to predict how the macroscopic hardness of a material will be affected by particle properties, solid fat content and the spatial distribution of the solid material. Processing conditions (heat and mass transfer) can affect some, or all, of these parameters. Quantitative relationships between network growth mode (rate of crystallization, type of growth, free energy of nucleation) and the resulting static structure, yield stress and Young modulus, as well as sensory texture will be explored in this presentation.