posted on 2016-03-15, 00:00authored byRobin Taylor, Jason C. Cole, Colin R. Groom
We describe an algorithm
for enumerating and classifying molecular
interaction motifs in crystal structures taken from the Cambridge
Structural Database. It was used to determine which motifs have the
highest propensity to occur in structures with Z′
> 1 and to show that these motifs also have a strong tendency to
occur
between symmetry-independent rather than symmetry-related molecules
in such structures. In noncentrosymmetric structures, they are predominately
stacking interactions and cyclic motifs containing strong or weak
hydrogen bonds. In centrosymmetric structures, OH···O
hydrogen bonds and C–H···π interactions
predominate. The proportions of centrosymmetric and noncentrosymmetric
structures with Z′ = 2 that might be explained
by the presence of one or more of these motifs were estimated. Motifs
with a strong tendency to form across inversion centers in structures
with Z′ = 1 were also identified (“inversion-favoring
motifs”). They are similar to those preferentially occurring
between symmetry-independent molecules in noncentrosymmetric structures
with Z′ = 2 but dissimilar to those in centrosymmetric structures.
However, a significant number of centrosymmetric structures with Z′ = 2 have inversion-favoring motifs both across
crystallographic inversion centers and between symmetry-independent
molecules. Several other results were obtained: for example, crystallization
with Z′ > 1 is more likely for molecules
with
1–3 hydrogen-bonding atoms.