Trans-(DL-isoleucinato-N,O)[tris(2-aminoethyl)amine-kappa(4)N]cobalt(III)diperchlorate

An exceptional 5:4 enantiomeric structure

The only crystals that could be grown from racemic solutions of the PF6−salt of the resolvable cation [Ru(2,9-dimethyl-1,10-phenanthroline)2(dipyrido[3,2-d:2′,3′-f]quinoxaline)]2+have translational symmetry only (space groupP1), contain nine independent sets of ions, and include numerous independent solvent molecules (11 acetone, one diethyl ether and possibly several water molecules). Layers of hydrophobic cations alternate with layers containing most of the anions and solvent molecules. All nine cations have the same basic conformation, which is distorted by the presence of the methyl substituents on the two 1,10-phenanthroline ligands. Four pairs of enantiomeric cations within a layer are related by approximate inversion centers; the ninth cation, which shows no sign of disorder, makes the layer chiral. Within the cation layers stripes parallel to [110] of six cations alternate with stripes of three; the local symmetry and the cation orientations are different in the two stripes. These stripes are reflected in the organization of the anion/solvent layer. Theca80:20 inversion twinning found indicates that enantiomeric preference is transmitted less perfectly across the anion/solvent layer than within the cation layer. The structure is exceptional in having nine independent formula units and an unbalanced set (ratio 4:5) of resolvable enantiomers. The difficulty in growing crystals of this material is consistent with its structural complexity.

Failures of fractional crystallization: ordered co-crystals of isomers and near isomers

A list of 270 structures of ordered co-crystals of isomers, near isomers and molecules that are almost the same has been compiled. Searches for structures containing isomers could be automated by the use of IUPAC International Chemical Identifier (InChI™) strings but searches for co-crystals of very similar molecules were more labor intensive. Compounds in which the heteromolecular A···B interactions are clearly better than the average of the homomolecular A···A and B···B interactions were excluded. The two largest structural classes found include co-crystals of configurational diastereomers and of quasienantiomers (or quasiracemates). These two groups overlap. There are 114 co-crystals of diastereomers and the same number of quasiracemates, with 71 structures being counted in both groups; together the groups account for 157 structures or 58% of the total. The large number of quasiracemates is strong evidence for inversion symmetry being very favorable for crystal packing. Co-crystallization of two diastereomers is especially likely if a 1,1 switch of a methyl group and an H atom, or of an inversion of a [2.2.1] or [2.2.2] cage, in one of the diastereomers would make the two molecules enantiomers.

A list of organic kryptoracemates

A list of 181 organic kryptoracemates has been compiled. This class of crystallographic oddities is made up of racemic compounds (i.e. pairs of resolvable enantiomers) that happen to crystallize in Sohnke space groups (i.e. groups that include only proper symmetry operations). Most (151) of the 181 structures could have crystallized as ordered structures in non-Sohnke groups. The remaining 30 structures do not fully meet this criterion but would have been classified as kryptoracemates by previous authors. Examples were found and checked with the aid of available software for searching the Cambridge Structural Database, for generating and comparing InChI strings, and for validating crystal structures. The pairs of enantiomers in the true kryptoracemates usually have very similar conformations; often the match is near-perfect. There is a pseudosymmetric relationship of the enantiomers in about 60% of the kryptoracemate structures, but the deviations from inversion or glide symmetry are usually quite easy to spot. Kryptoracemates were found to account for 0.1% of all organic structures containing either a racemic compound, a meso molecule, or some other achiral molecule. The centroid of a pair of enantiomers is more likely (99.9% versus 99% probability) to be located on an inversion center than is the centroid of a potentially centrosymmetric molecule.