E0015

What is a Routine Crystal Structure Determination? A.L.Spek, Utrecht University, 3584 CH Utrecht, The Netherlands

Really routine crystal structures can be determined nowadays by experts on a one-hour basis or often less, given access to a state-of-the-art CCD-detector based diffractometer system. For one, we have solved structures from 10 minute KappaCCD/rotating anode datasets.
However, contrary to what some vendors claim, the real world problem is that many samples made available for structural analysis are far from standard; at least not within our National facility for single crystal structure determination (e.g. poorly reflecting, small, twinned, disorder, losing solvent etc.).
Fact is that there are many pitfalls on the road to a reliable structure determination suitable for publication. The catch is that one needs to be an experienced crystallographer to recognize possible trouble.
Unfortunately, the literature is contaminated with publications of false structural results due to the fact that all warning signals were ignored.
Disorder can be real or an artifact. Atoms can be assigned the wrong element type. The space group assigned may be wrong. Essential hydrogen atoms may have been missed resulting in false conclusions about the chemistry.
Tools are now under development aiming at the automated analysis of the results of a crystal structure determination (at least generating warning signals).
A large number of tests have now been implemented in the checking/validation software that is run routinely for structures submitted for publication in Acta Cryst.
This talk will present a few interesting non-standard structures/problems and their solution from the over 2000 structures we studied.
Most of our experience with problems encountered is 'archived' in a program named PLATON. (see http://www.cryst.chem.uu.nl/platon).