Powder diffraction data rarely provide all the structural detail a crystallographer needs. But, if theoretical calculations in the form of dispersion-corrected density functional theory (DFT-D) can validate the data, then for structures where sufficiently large crystals are inaccessible the picture can be made clearer.
In 2010, Jacco van de Streek and Marcus Neumann of the University of Copenhagen, Denmark and Avant-garde-Materials Simulation, in Germany, tested the concept against data sets from 225 high-quality single-crystal determinations of organic molecules as a benchmark. The team reported that DFT-D offers a good approximation for those molecules, Now, the team has verified that the approach works with powder data on 215 sample organic structures each published in an IUCr journal identified through a search of the Cambridge Structural Database [van de Streek, J. & Neumann, M. A. (2014), Acta Cryst. B70, 1020-1032; doi: 10.1107/S2052520614022902].
The team points out that density functional theory techniques of all the approaches to quantum mechanical calculations for molecules offer a favorable trade-off between accuracy and speed.
They report that three of the 2010 batch of 225 published single crystal structures were found to be incorrect. In contrast, almost 9 percent (19 of the 215 powder structures) were "demonstrably in error". The team does reassure us that the majority of those erroneous structures were only incorrect in minor ways - minor space-group revisions, exchanges of atoms with similar electron densities such as a carbon for a nitrogen, and ambiguities involving hydrogen atoms, for example where hydrogen bonding was involved. They add that some ambiguities were reported by the authors of the original papers but those authors did not have available to them the tools to remedy this situation.
van de Streek and Neumann have now demonstrated that not only can Dispersion-corrected Density Functional Theory (DFT-D) calculations provide "an independent source of structural information about organic crystal structures" to help improve the structures obtained, they can corroborate or counter claims for a given structure. Moreover, this and the 2010 study offer a warning to users of powder diffraction analysis regarding an often hard to detect phenomenon called "preferred orientation", in which the measured data are biased and do not accurately reflect the underlying crystal structure. The team found that the most troublesome ambiguities were manifest in those studies in which preferred orientation was present. This phenomenon should, the team suggests, be treated with greater suspicion than is currently the case when reporting a "definitive" crystal structure.
Talking to the IUCr, van de Streek told us "The main limitation of the DFT-D approach is that only static structures can be addressed and we are currently investigating methods to include the effects of temperature such as thermal expansion and disorder".