The trouble with crystal polymorphism

By Alejandro Marangoni

In This Section

June 2011

Editor's Note: The following article is based on a Hot Topics: Tough Topics to Teach presentation, "Everything You Wanted to Know About Lipid Polymorphism, but Were Afraid to Ask," at the 102nd AOCS Annual Meeting & Expo (May 1-4, 2011) in Cincinnati, Ohio, USA.

Crystal polymorphism is one of the most widely studied structural characteristics of fats used in spreads, shortening, and confectionery applications. Simply put, polymorphism refers to the fact that triacylglycerols (TAG), and other lipids, can crystallize in different crystal types-all having, on average, the same chemical composition. Each polymorph will have a characteristic melting temperature, powder X-ray diffraction pattern, infrared spectral signature, or Raman spectra. At least three main polymorphic forms have been identified, and are ever-present in "fat lore." These include the alpha, beta prime, beta forms (listed in order of increasing stability), melting point, and packing density. Good-quality chocolate is associated with its content of cocoa butter in the beta, form V, polymorph, while "good shortening" or "good spreads" are associated with the presence of fat crystallized in the beta prime form. Even though this relationship is very poorly understood and many times is not verifiable, we still like to use statements such as these.

When they crystallize, TAG pack in very specific structural arrangements with characteristic symmetries. The symmetry properties of a crystal are given by the type of unit cell present in such crystal. The packing arrangement of atoms within the crystal defines the type of unit cell structure present in the crystal (Fig. 1).

Fig. 1 shows the packing arrangement of the beta prime polymorph of 1,2-dipalmitoyl-3-myristoyl-sn glycerol determined by Professor Kiyotaka Sato's group from Hiroshima University (J. Lipid Res. 42:338-345, 2001). Sato is a past recipient of the prestigious AOCS Stephen S. Chang Award. Indicated are the long axis (c) of the unit cells and the two small axes (a and b). Notice that two TAG molecules are stacked on top of each other in the long axis direction. This is often where the trouble begins.

We like to talk about what type of "unit cell" is present in the different polymorphs. However, we do not do this; we resort to the concept of a "subcell," some imaginary sub-structure within the fatty acid chains of the TAG. Thus, in this way, we consider TAG to be a collection of alkanes. The alpha form is supposed to have a hexagonal unit cell, the beta prime is supposed to be orthorhombic, while the beta form is supposed to be triclinic.

However, close inspection of Sato's beta prime structure quickly reveals that the unit cell is monoclinic, but we call this polymorph orthorhombic. How can this be? We have the same problem in all single-crystal structures for the beta form-they are not triclinic! So, we are left with this concept of the subcell. This concept has many problems, including that it is not possible to define the type of subcell present experimentally due to a lack of sufficient information that can be gathered from powder X-ray diffraction experiments. For example, Fig. 2 shows the characteristic powder X-ray diffraction patterns for the three different polymorphic forms of fully hydrogenated canola oil.

In the wide angle region (small spacings), we have only one peak for the alpha and beta forms and two for the beta prime forms. We need about 20 peaks to carry out good indexing. So how do we know these structures are what we think they are?

If the food chemistry instructor wants to delve into structural discussions about the different polymorphs, it is necessary to take another approach. In short, we need to understand the concept of a unit cell of a TAG, not of a subcell of a TAG. We need to understand that a crystal is made up of many crystal planes formed by the regular arrangement of atoms within the crystal. These planes can be characterized by their Miller indices. Ultimately, the position of a powder X-ray diffraction "peak" is defined by the inter-planar distance for one of the many planes present in the crystal, while the relative intensity of the peaks is given by the structure factor. Here we will give a brief peek at some of the work carried out by Stefan Idziak at the University of Waterloo (Ontario, Canada) and David Pink (St. Francis Xavier University, Antigonish, Canada) in collaboration with Alejandro Marangoni at the University of Guelph (Ontario), trying to define why we obtain these characteristic powder X-ray spectra for the different polymorphs. It is intriguing that fats with very dissimilar TAG compositions all, more or less, give the same characteristic diffraction patterns for the different polymorphs. Here we will show how this is only a function of the enormous asymmetry of the TAG molecules' unit cell of the TAG molecules-a huge long axis and two very small short axes.

In the end, however, we can/should go back to the basic concept of polymorphism, refer to the differences in melting points and stability, and discuss the importance of the crystal type to fat functional properties. One may want to take this concept further and qualify the statements about the importance of polymorphism on functional properties. Polymorphism is important since it defines the melting properties of the material, but the nanostructure and microstructure of the material are equally or more important (Fig. 3), and not necessarily always linked to a particular polymorphic state. One can have a very small beta crystal or a huge alpha spherulite.

Alejandro Marangoni is a professor in the Department of Food Science, University of Guelph (Guelph, Ontario, Canada). He may be contacted at