Abstract
In the absence of single crystals, powder diffraction data (especially synchrotron data) can provide a means of determining the crystal structure of an active pharmaceutical ingredient. The process of solving and refining a crystal structure using powder data is summarized, and suggestions for overcoming challenges are given for the various steps. Indexing can often be a bottleneck. Measures of the quality of a Rietveld refinement - including statistical, graphical, and chemical reasonableness (particularly the root-mean-square displacement between a Rietveld-refined and a DFT-optimized structure) - are described, and examples are given of what to expect, based on a large number of structures determined using synchrotron powder diffraction data. Recent crystal structures of important and/or interesting molecules are reviewed. Examples illustrating the accuracy which can be achieved, interesting features of the process of structure solution, the distinction between salts and co-crystals, and phase transitions are also reviewed.