Journal of the American Chemical Society, Vol.116, No.2, 523-527, 1994
Molecular-Crystal Engineering by Shape Mimicry
Based on the high statistical preference for organics in general and true racemates in particular to pack in one of the centrosymmetric space groups, formation of molecular crystals of single enantiomers of shape-symmetric molecules was predicted to occur with approximate crystallographic symmetry. Thus, enantiomerically pure sulfoxide 1 was designed with two segments of nearly identical shape but with significantly different electron-donor/acceptor properties. Packing of 1 was predicted to occur in a near-centrosymmetric or quasicentrosymmetric fashion, with pairs of 1 arranged as would be expected for a true racemic crystal. Such a packing motif would lack true centrosymmetry and, in certain space groups, could result in a polar direction in the crystal and net additivity of the vectors from nitrogen to sulfur (the direction of polarizability for the molecules). Enantiomeric sulfoxide 1 does form molecular crystals with near centrosymmetry, mimicking P2(1)/c, where the vector from the dimethylamino group to sulfur of one molecule is antiparallel to that from the isopropenyl group to the sulfur of the other. Thus, there is a substantial net directionality of polarizability vectors from nitrogen to sulfur in this pair. By contrast, the sulfoxides 2 and 3, wherein the two phenyl [GRAPHICS] moieties in each differ significantly in shape, form molecular crystals in which the vectors from nitrogen to sulfur are antiparallel and there is substantial net additivity of the dipole vectors from sulfur to oxygen. Crystal data for 1 at -80-degrees-C : monoclinic, P2(1) (No. 4), Z = 4 (quasi-P2(1)/c), a = 7.6716(13) angstrom, b = 8.2498(20) angstrom, c = 26.1127(49) angstrom, beta = 114.716(13)degrees, for data with I > 2sigma(I), R1 = 0.0505, wR2 = 0.1059. Crystal data for 2 at 25-degrees-C : monoclinic, P2, (No. 4), Z = 2, a = 7.8030(6) angstrom, b = 6.0355(6) angstrom, c = 17.037(2) angstrom, beta = 96.899(7)degrees, for data with I > 2sigma(I), R1 = 0.0542, wR2 = 0.1353. Crystal data for 3 at -80-degrees-C : monoclinic, P2(1)(No.4), Z = 2, a = 7.7523(8)angstrom, b = 5.9869(7) angstrom, c = 14.8133(16) angstrom, beta = 103.244(8)degrees, for data with I > 2sigma(I), R1 = 0.0383, wR2 = 0.1041. All three structures were refined by full-matrix least-squares methods versus Absolute value of F2.??-c?