On the inclusion of solvent molecules in the crystal structures of organic compounds

The Cambridge Structural Database has been searched for all crystal structures including organic solvent molecules (solvates) and solvent water molecules (hydrates). Well above 300 different solvent molecules were identified and the frequencies with which they occur in crystal structures, as a function of the year of publication, were established. The crystal structures are classified as 'organic' and 'metalloorganic'; it is shown that the relative prevalences of various cocrystallized solvents are different in the two groups. Several frequently used organic solvents are also common ligands for metal ions. Special interest has been focused on the existence of heterosolvates, i.e. crystal structures which include more than one type of solvent molecule. Up to five different types of solvent molecules were found in a single crystal structure. It is suggested that the use of solvent mixtures during crystallizations may prove to be a more useful and versatile approach for obtaining crystals of high-molecular-weight organic compounds than has hitherto been recognized.

Symmetry, pseudosymmetry and packing disorder in the alcohol solvates of L-leucyl-L-valine.

The dipeptide L-Leu-L-Val can crystallize as a hydrate in the hexagonal space group P6(5) [Görbitz & Gundersen (1996). Acta Chem. Scand. 50, 537-543], but forms 1:1 solvates when methanol, ethanol and 2-propanol are used as precipitating agents. The structures of these complexes can be divided into hydrophobic and hydrophilic layers. The alcohol molecules are embedded in the hydrophobic layers, but with the hydroxyl groups as crucial parts of the hydrogen-bonding network. L-Leucyl-L-valine-methanol (1/1) crystallizes in the space group P2(1) with Z = 2. L-Leucyl-L-valine-ethanol (1/1) has very special systematic absences, and the structure could not be solved with direct methods. Unraveling the actual build-up of the crystal was an unusual process involving modeling with molecular graphics programs. The solution shows a structure which formally belongs to the space group P2(1), with four dipeptide molecules and four solvent molecules in the asymmetric unit (Z = 8). Additional non-crystallographic symmetry in three dimensions fixes all cell angles, including beta, to 90 degrees. L-Leucyl-L-valine-2-propanol (1/1) is structurally closely related to the ethanol solvate, but owing to a rare type of packing disorder the length of the a axis is halved (Z = 4, P2(1)2(1)2(1)). The hydrogen-bonding pattern is still the same as in the ethanol solvate, which means that the hydrogen-bond periodicity along the a axis in the 2-propanol solvate is two unit-cell lengths.

Molecular aggregation in crystalline 1:1 complexes of hydrophobic D- and L-amino acids. I. The L-isoleucine series.

The amino acid L-isoleucine has been cocrystallized with seven selected D-amino acids including D-methionine [L-isoleucine-D-methionine (1/1), C(6)H(13)NO(2).C(5)H(11)NO(2)S, amino-acid side chain R = -CH(2)-CH(2)-S-CH(3)] and a homologous series from D-alanine [L-isoleucine-D-alanine (1/1), C(6)H(13)NO(2).C(3)H(7)NO(2), R = -CH(3)] through D-alpha-aminobutyric acid [L-isoleucine-D-alpha-aminobutyric acid (1/1), C(6)H(13)NO(2).C(4)H(9)NO(2), R = -CH(2)-CH(3)] and D-norvaline [L-isoleucine-D-norvaline (1/1), C(6)H(13)NO(2).C(5)H(11)NO(2), R = -CH(2)-CH(2)-CH(3)] to D-norleucine [L-isoleucine-D-norleucine (1/1), C(6)H(13)NO(2).C(6)H(13)NO(2), R = -CH(2)-CH(2)-CH(2)-CH(3)] with linear side chains, and D-valine [L-isoleucine-D-valine (1/1), C(6)H(13)NO(2).C(5)H(11)NO(2), R = -CH-(CH(3))(2)] and D-leucine [L-isoleucine-D-leucine (1/1), C(6)H(13)NO(2).C(6)H(13)NO(2), R = -CH(2)-CH-(CH(3))(2)] with branched side chains. All the crystal structures are divided into distinct hydrophilic and hydrophobic layers. In the five complexes with amino acids with linear side chains the polar parts of the D- and L-amino acids are related by pseudo-glide-plane symmetry, and four of them have remarkably similar molecular arrangements. The D-valine and D-leucine complexes, on the other hand, are structurally quite different with the polar parts of the D- and L-amino acids related by pseudo-inversion. Differences in the hydrogen-bond pattern in the two molecular arrangements are discussed.

What is the best crystal size for collection of X-ray data? Refinement of the structure of glycyl-L-serine based on data from a very large crystal.

The dipeptide Gly-L-Ser was crystallized as part of a study on hydrogen-bonding patterns in the structures of dipeptides. Hydrogen-bond donors and acceptors have been assigned ranks (1 is best, 2 is next best etc.), and the observed hydrogen-bond connectivity is compared with the hypothetical pattern in which the rank n donor associates with the rank n acceptor (n = 1, 2,.), and with the pattern observed in the retroanalogue L-Ser-Gly, which contains the same functional groups. Crystallization of the title compound produced very bulky crystals. Rather than reducing the size of one of these before data collection, three data sets with different exposure times were collected with a Siemens SMART CCD diffractometer on a very large specimen (2.2 x 2.0 x 0.8 mm). The crystal was subsequently shaped into a 0.30 mm-diameter sphere for collection of two additional data sets. The discussion of the refinement results focus on the effect of absorption correction for the various data sets, and a comparison of geometrical and thermal parameters. One advantage of using a large crystal, the great speed with which data can be obtained, has been exemplified by collection of a complete data set of good quality in less than 25 min.