On the origin of topological differences between experimental and theoretical crystal charge densities

Topological analysis of experimental and theoretical (molecular and crystal) electron densities of p-nitroaniline and p-amino-p'-nitrobiphenyl reveals considerable discrepancies between experiment and theory for the bond critical points properties. Particularly large differences occur for the positive curvature along the bond path (lambda(3)). The differences become somewhat smaller when more extended basis sets and correlation effects are introduced in the theoretical calculations. The effect of the crystal matrix on the properties of bond critical points is evaluated for the p-nitroaniline molecule using the 6-21G** and 6-31G** basis sets. The differences between the isolated molecule and the molecule in the crystal are too small to explain the quantitative disagreement between the theoretical and experimental topologies reported in the literature and found in the current study. For most bonds, the observed changes in the properties of the electron density agree well for both basis sets but some discrepancies are found for changes in lambda(3) for N-H and aromatic C-C bonds. When the theoretical densities are projected into the multipole density functions through refinement of the theoretical structure factors, the topological properties change and differences between theory and experiment are reduced. The main origin of the observed discrepancies is attributed to the nature of the radial functions in the experimental multipole model.