Challenges and approaches to crystal structure determination
PDF Print E-mail

Solid state research, such as polymorph investigations, salt or co-crystal screening, benefits from detailed understanding and knowledge of the crystal structure.  The information gained from the intra- and intermolecular interactions, the density of the crystal, the packing index and the size and characteristics of voids is fundamental.  The knowledge is used to select the thermodynamic most stable polymorph, predict how the molecule will behave under different humidity levels and provide information of the best counter ions for salt formation or predict co-crystal properties.

There are several techniques that provide structural information on the solid state appearance of an API such as ss-NMR, IR or Raman.  However, none of these techniques beat the power of single crystal X-ray diffraction (SC-XRD).  Only SC-XRD will give the full understanding on the molecular build of the API in the solid state.

Picture 1: Packing of {p-[2-Hydroxy-3-(isopropyl-λ4-ammonium)propoxy]phenyl}acetamide Cinnamate

Unfortunately, it is not always possible to get single crystals of sufficient quality to perform single crystal analysis.  Furthermore, it is often difficult to obtain crystals of metastable forms.  In addition, during the time it takes to grow the single crystals phase transformation can take place to more stable forms.  An affordable alternative to SC-XRD is structure determination from powder data.  This approach applies for single molecule crystals, as well as, for multi-component crystals such as salts, co-crystals, hydrates and solvates.

Determining the crystal structure from powder is not trivial and requires experienced scientists to be successful.   It starts with the preparation of the powder that needs to finely grinded without affecting the crystallinity of the sample.  In a well prepared sample the preferred orientation effect on the peak intensities will be minimal.  The crystallinity of the sample should be high such that the FWHM (Full Width at Half Maximum) at low 2q reflections should not exceed 0.06°.  The signal to noise ratio should allow for miniscule reflections to be visible.

Picture 2: Selected torsion angles for rotating in the molecule of 6-Methoxycarbonyl-2-Naphtoic Acid

Once an acceptable diffraction patterns is recorded and carefully indexed the calculations are started to obtain the correct space group.  Selection of the possible space groups out of the 230 total available space groups is based on molecular symmetry and chemical knowledge.  If the molecule crystallizes in a higher symmetry than monoclinic, the powder pattern will fall into multiple space groups which all have to be considered for the final structure.  For refinement of the cell parameters the Whole Powder Pattern Decomposition method is used.  Subsequently a rough model is build based on the 2D molecular structure and acceptable density (from thermal data).   This preliminary model is further optimized by simulated annealing.  In this process the molecule is moving around in the cell and functional moieties are rotated around selected torsion angles.  Based on experience, H-bonds schemes, intra- and intermolecular bonds several acceptable models are chosen.  Finally, the calculated powder patterns are compared to the experimental XRPD (Rietveld analysis) and the model that best explains the crystal structure is selected.