Prediction of crystal structures of two monosaccharides by potential energy minimisation

On reliability of Rwp in structure modelling

In our previous paper on prediction of crystal structures of two monosaccharides we pointed to a contrast between rather unrealistic potential energy values paired with acceptable Rwp factors. By comparing the energy values E(potential) to E(HFSCF) and by the refinements combining trial structures and |F0| from a single crystal it was found that Rwp could be a misleading indicator in structure prediction based on powder data.

Prediction of crystal structure of a rare monosaccharide – influence of model geometry

Crystal structure of a hexagonal monosaccharide, C11H18N2O4, was succesfully predicted by potential energy minimisation. Four starting molecular models differing in ring conformations were based on similar structures and the positions of two substituents (-NH2, -CN) were found by semiempirical AM1 calculations. Generated structures were ranked according to both potential energy values and corre-sponding Rwp ´s. Calculated potential energy was sensitive to changes in the geometries of the sachcharide rings as well as to the positions of substituents. Rigid body Rietveld refinements did not improve accuracy of the crystal structures with low Rwp ´s and high potential energy.

Decafluoroquarterphenyl - crystal and molecular structure solved from X-ray powder data

The structure of decafluoroquarterphenyl has been solved and refined from laboratory X-ray powder data. Structure analysis and ab initio Hartree-Fock calculation showed that the molecules are held by electrostatic interactions and weak hydrogen bonds. A Rietveld refinement supported by slack constraints gave dihedral angles which agreed with those found in similar compounds. In molecules such as these, where the planarity of the rings is known a-priori, and there is little improvement in the fit to the diffraction data at the expense of distorting the structure, conventional slack constrained Rietveld refinement may not bring much additional information and therefore may not be worth the additional effort at the end of the structure solution process.