[pullquote]
If you have a crystallizer or are planning to install a crystallization process, then you know the importance of solubility curves. Yet, it’s amazing how many plants run without such curves or use ones that are ancient, perhaps many decades old. Empirical data underpinned the development of these old-time processes, which probably will run forever without any changes — at least we hope so. However, hope isn’t a strategy or a plan. Something such as a raw material will change or customers will want a different particle size distribution.
This is when I show up asking for data such as a solubility curve, and pose a lot of questions about the crystals and the process. Even plants that have a curve are stymied when I ask about potential crystal phases such as polymorphs or solvates. The concern I have for crystal phase isn’t limited to the crystallization process; this physical property may be important in drying, transport and storage of the chemical.
I can’t emphasize enough the need to look very hard for the existence of these phases along with any temperature sensitivity. Any heat flux that shows up during temperature changes is a concern. We were developing a new pharmaceutical and in the final stages of getting U.S. Food and Drug Administration approval when the solubility of a batch changed dramatically; we even saw a slight shift in the x-ray pattern. When I asked about polymorphs, the team responded that it didn’t think any existed. However, one chemist said she may have seen one. Molecular modeling identified six potential polymorphs; two were robust. The problem was resolved by avoiding a temperature region that allowed one of these polymorphs to form.
I’ve mentioned phase in previous columns (e.g., “Get a Solubility Curve”). Here, let’s look at some tools for identifying polymorphs and solvates to help you avoid or address production problems, or allow you to produce a more desirable product:
Molecular modeling. This is a theoretical method and computational technique that attempts to mimic the behavior of molecules. Many advances have occurred in this technology. Hand calculations can define simple systems. The common feature of molecular modeling is the atomic level description of the molecule. You can treat atoms using molecular mechanics or by explicitly modeling electrons using a quantum chemistry approach. I ran some rather crude computer programs over 20 years ago to identify the polymorphs in the pharmaceutical problem cited above.
X-ray diffraction. This method will identify structural differences of a material. By measuring the angles and intensities of a diffracted beam, a chemist can produce a three-dimensional picture of the crystal. It’s important to look at material produced under different conditions to identify the most stable crystal form and any polymorphs.
Differential scanning calorimetry. Many variations of this technique exist but the basic idea is to identify phase changes as a sample is heated. The amount of energy absorbed or released may indicate a polymorph or solvate.
Raman and infrared spectroscopy. Raman spectroscopy is a technique to observe vibrational, rotational and other low-frequency modes of a chemical. It can identify molecules that have a slight variation in structure. Infrared is another method that can serve for quantitative analyses of unknown substances or to determine the structural properties of known substances.
Microscopy. It’s amazing what you can learn from looking at your product under an optical microscope. However, a scanning electron microscope or a transmission electron microscope will reveal differences in the surface or composition of the sample and help identify structural distinctions that indicate the presence of different chemicals.
Drying rate or critical moisture. Often solvates or polymorphs dry differently due to altered heat capacity caused by their distinct structure.
The most important conclusion you can reach by using these methods is the identification of the most stable form of the chemical. If that form is the most effective product (e.g., the drug that works the best), the additional information about when an unstable form occurs will allow you to always make the desired product. If an unstable form is more effective, then these data will allow you to tailor your process to avoid the stable product.
