Testing theory beyond molecular structure: Electron density distributions of complex molecules



The electron density distribution (EDD) of a molecular system can be determined experimentally from elaborate X-ray diffraction measurements or calculated with quantum mechanical methods: This provides a unique opportunity for mutual validation of the experimental and theoretical methods—a validation that goes far beyond comparison of molecular structures. Two examples of complex molecular systems of biologic relevance are presented. The first is the cocrystallized complex of betaine, imidazole, and picric acid, 1, which is a 75-atom molecular complex serving as a model for the active site in the serine proteases class of enzymes, the so-called catalytic triad. For 1 the experimental charge density was determined by combined modeling of single crystal synchrotron X-ray and neutron diffraction data measured at 28(1) K, and it is compared with ab initio theoretical calculations at the B3LYP/6-311G(d,p) level of theory. Overall, the agreement is good, but in one strong N[BOND]H[BOND]O hydrogen bond clear differences are observed. The second example concerns the EDD of the mixed valence trinuclear oxo-centered iron carboxylate, [Fe3O(OOCC(CH3)3)6(NC5H5)3], 2. This molecule contains 133 atoms (542 electrons) including three open-shell iron atoms, and the experimental investigation is based on synchrotron X-ray diffraction data. Calculations in the experimental geometry at the commonly used UB3LYP/LanL2DZ level of theory are not able to reproduce a number of experimentally observed electron density features. In particular, the sp3-like distribution on the central oxygen atom and the electron deformations on the iron centers are at variance with experiment. © 2003 Wiley Periodicals, Inc. Int J Quantum Chem, 2004